Accepted Poster

Last Update : 2023/08/13 14:34

The place for all poster presentation is C (Build.15), B1.
It is open space, so you cannot miss it.

In the space, you can find the space ID (C**) on the board.
Please find your space from this page.

List of all posters, Aug. 22, Aug. 23, Aug. 24, Aug. 25.

[10939] Exploring Molecular Machine Learning Models for Activity-Cliff Prediction

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Pairs of almost identical molecules that exhibit a large activity difference against a given biological target are called activity cliffs and form an important source of pharmacological information. We computationally investigate the capabilities of current machine-learning-based activity-prediction models to detect activity cliffs. The models are weak at identifying activity cliffs unless the activity of one molecule in the pair is known. Increasing activity-cliff sensitivity might form a research path towards further improving computational activity-prediction.
• Author(s) : Markus Dablander, Thierry Hanser, Renaud Lambiotte, Garrett M. Morris

[10941] Existence Results for an eigenvalue Riesz-Caputo Fractional Boundary Value Problem

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : In this work, we study the solvability of an eigenvalue fractional boundary value problem depending on the Riesz-Caputo derivative. By using Green’s function properties, we provide the existence of solution via Schaefer’s fixed point theorem. We end this work by presenting a Lyapunov-type inequality and a bound for the possible eigenvalues for the corresponding problem.
• Author(s) : Rabah Khaldi, Assia Guezane-Lakoud

[10942] A model-averaged approach of concordance correlation coefficients for longitudinal overdispersed Poisson data

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Variance components (VC) is an approach to estimate concordance correlation coefficient (CCC) through adjusting for covariates. To avoid fitting data with a misspecified model, corrected conditional Akaike information criterion (CCAIC) and corrected conditional Bayesian information criterion (CCBIC) measures are adopted for model selection in Poisson mixed-effects models. This study focuses on proposing the model-averaged approach by combining the CCC estimators of VC with model selection via CCAIC and CCBIC for longitudinal overdispersed Poisson data.
• Author(s) : Miao-Yu, Tsai

[10949] RICE HUSK IN THE HYDROCARBONS INDUSTRY

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Hydrophobic substances have multiple applications in industrial uses. We are interested in describing the behavior of rice husks with hydrocarbons. To do it, we
  use the equations of isotherms de Freundlich and Langmuir to analize the rice husk
  adsorption for hydrocarbons.
  
• Author(s) : David Matheo Vargas Huertas, Javier Esteban Martinez Caldas, Juan Jose Camargo Carbonell, Juan Steban Garzón Trujillo, María Isabel Romero Rodríguez, Jorge Eliécer Carillo Velásquez

[10950] Parameter identification analysis for incompressible viscous flow with interface

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In this study, we present the parameter identification analysis for incompressible viscous flow with interface. A dam break model is introduced as the numerical model, the identification analysis is performed based on the adjoint variable method. The incompressible viscous flow analysis with interface is carried out based on the characteristic finite element method using the semi-Lagrange method, and the volume of fluid method is introduced to analyze the flow behavior of two fluids.
• Author(s) : Masaya Kobayashi, Takahiko Kurahashi, Toshiaki Kenchi, Toshihiko Eto

[10951] Multi stage deep reinforcement learning for Automatic three-dimensional cephalometric landmark detection

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Landmark detection of 3D CT data for cephalometry in orthognathic surgery is time-consuming and error-prone. We here propose automatic 3D cephalometric annotation system based on multi-stage deep reinforcement learning (DRL) and volume-rendered imaging without expert intervention. This system considers geometrical characteristics of landmarks and simulates the sequential decision process underlying human professional landmarking patterns. We also provide the results of applying the proposed algorithm to actual clinical patients with deformities.
• Author(s) : Sung Ho Kang, Kiwan Jeon, Sang-Hoon Kang, Sang-Hwy Lee

[10954] A lower-order weighted least-squares finite element method for poroelasticity problems in rheology

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The work concerns the behavior of the approximate solution of Biot’s consolidation problem by the least-squares finite element method (LSFEM). In the case of fluid flow in deformable porous media, these consist of fluid pressure and flux as well as displacement field and stress tensor. We consider a stabilized weight in the LSFEM with lower-order finite element space and illustrate the method’s performance. Further, we extend the LSFEM to physical problems.
• Author(s) : Hsueh-Chen Lee, Hyesuk Lee

[10956] Poincaré Section for Hide Coupled Dynamo Model,

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Abstract: Poincaré surface of the section is an important tool in the dynamical system which allow us to imagine and understand the numerical solution behavior of the system. Hide’s couple dynamo model is rich and worth studying. In this paper, we apply the Poincaré surface of section in three cases periodic orbit motion, regular motion, and chaos motion
• Author(s) : Ali Allahem

[10957] Identification analysis of defect topologies by self-attention-based machine learning (Effect of number of training data on identification accuracy)

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In this study, we present the identification analysis of defect topologies in the concrete structure by the self-attention-based machine learning. In this analysis, the scalogram is generated by the hammering response data based on the continuous wavelet transform, and is used as the learning data. The computation of the self-attention and the transition is repeated in the self-attention network, and weighting parameters which are used to estimate the defect topologies can be obtained
• Author(s) : Kazuki Yamamoto, Takahiko Kurahashi, Yuki Murakami, Fujio Ikeda and Ikuo Ihara

[10959] Identification analysis of defect topologies using level-set-based topology optimization with weighted sensitivity

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In this study, we present the defect topology identification analysis based on the level-set-based topology optimization. In this analysis, the hammering response data is employed to identify the defect topologies. The equation of motion in three dimensions is employed to simulate the oscillation behavior in the structure. The weighted sensitivity method is applied to identify the defect topologies, and some numerical results are shown in this study.
• Author(s) : Towa Koike, Takahiko Kurahashi,Masayuki Kishida, Yuki Murakami and Fujio Ikeda

[10961] Mathematical Modeling for a Bioglass Bioactivity Degradation

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In this work we develop a mathematical model, to analyze the dissolution and bioactivity for a bio-glass. The development of porous bioactive glasses is part of a multidisciplinary task. For example in pharmacology, the developments of bioactive systems are suitable for many applications such as prolonged-release of drugs. We first derive a model based on a reaction diffusion system; then we propose a numerical framework to show useful results in real applications.
• Author(s) : Aymen Hadji, Fatma Zohra Nouri

[10966] Simulation of Synchronization with Neuronal Population Firing Model

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Understanding the human brain, novel simulation of synchronization with neuronal population firing model is developed. Here, the simulation is consisted with two scale model. The one is micro model that synchronization with neuronal spiking model in a single neuronal population. The other is macro model that neuronal population firing model that is taking account micro model. With the simulation we can find several synchronization dynamics of brain. This work contributes to whole brain simulation soon.
• Author(s) : Sho IKEDA, Toshiaki ITOH

[10969] Epidemiological Modeling of Health Information Dynamics on Twitter

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Epidemiological models are used to understand how information spreads on Twitter, dividing individuals (or users) into groups and simulating their interaction. In TwitHComm, we found that the tweet data obtained from @DOHgovph does not achieve an epidemic state whereas @WHO does. In TwitHCommS, despite increasing the number of positive sentiment tweets, users on Twitter are influenced by negative sentiments caused by the greater rate of negative sentiments among the users.
• Author(s) : Feeroz R. Yusoph, Angelyn R. Lao

[10970] On p(t)-Laplacian fractional differential equations

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We study the existence of solutions for a class of differential equations involving mixed type fractional Caputo derivatives and the p(t)-Laplacian operator which is a non-classical growth operator and arise from various field of sciences such image restoration, elasticity theory, electrorheological fluid,….
  By means of some fixed point theorems, we prove the existence of solutions for the considered problem.
  
• Author(s) : Assia Guezane Lakoud, Rabah Khaldi

[10971] Systems biology approach to understanding azole resistance mechanisms in Candida albicans

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The significant increase in fluconazole-resistant Candida albicans calls for a need to search for other possible drug targets. In this study, we constructed a mathematical model, based on the data collected from the literature, of the ergosterol biosynthesis pathway in C. albicans. Interestingly, we found an increase in the susceptibility of C. albicans to fluconazole with increasing concentrations of the sterol-methyltransferase enzyme, making it a potential drug target as an adjunct to fluconazole.
• Author(s) : Paul K. Yu, Llewelyn S. Moron-Espiritu, Angelyn R. Lao

[10972] A multiscale model for axon durotaxis

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : During development, axons extend from neurons and navigate through the extracellular environment in search of predetermined targets. Extracellular signals inform the growth trajectory of axons. One such signal is spatial variation in substrate stiffness.We formulate a multi-scale mechanical model of the axon which incorporates the axon-substrate interaction. We simulate growing axons in environments of spatially varying stiffness and show that such variations can play a significant role in the formation of axonal growth patterns.
• Author(s) : Christoforos Kassianides, Hadrien Oliveri, Alain Goriely

[10974] Structure-Preserving Neural Networks for Hamiltonian Systems

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : When solving Hamiltonian systems using numerical integrators, preserving the symplectic structure is crucial. We analyze whether the same is true if neural networks (NN) are used. In order to include the symplectic structure in the NN’s topology we formulate a generalized framework for two well-known NN topologies and discover a novel topology outperforming all others. We find that symplectic NNs generalize better and give more accurate long-term predictions than physics-unaware NNs.
• Author(s) : Philipp Horn, Barry Koren

[10980] Similarity and Finite Difference Solution on Biomagnetic Flow and Heat Transfer of Blood-Fe3O4 through a Thin Needle

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : A magnetic fluid is composed of a base fluid and magnetic particles, where magnetic particles are carefully distraught in the base fluid. Here, we will assume that blood is the base fluid that exhibits electrical conductivity and polarization properties and Fe3O4 as magnetic particles. The addition of Fe3O4 into the blood can remarkably ameliorate the properties of the blood’s thermal conductivity. Such physical aspects can play a vital role in biomedical and bioengineering. The presented
• Author(s) : Abdulaziz Alsenafi, Mohammed Ferdows

[10981] The Foldy–Lax approximation is valid for nearly resonating frequencies

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Waves propagating in the presence of a cluster of inhomogeneities undergo multiple interactions. When these inhomogeneities have sub-wavelength sizes, the dominating field due to these multiple interactions is the Foldy-Lax-Field. The question is whether we can reconstruct this Foldy-Lax-Field from the scattered field measured far from the inhomogeneities cluster. We will show that exciting the cluster by incident frequencies which are close to the real parts of these resonances will reconstruct the Foldy-Lax-Field.
• Author(s) : Abdulaziz Alsenafi, Ahcene Ghandriche, Mourad Sini

[10982] Mathematical and Numerical Study of a Stem Cell Problem

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The multiphase flow is used to refer to any fluid flow consisting of more than one phase or component. In fluid mechanics/dynamics, it is a simultaneous flow of materials with different states/phases. This work aims the need to model and predict those flows behaviour and their related manifest phenomena. More precisely, here we present a fundamental understanding of a stem cell problem in orthopaedic tissues, for which we illustrate a mathematical model and numerical results.
• Author(s) : Ibtissem Hadji, Fatma-Z Nouri

[10983] Compartment Models for Ideas on Social Media Networks

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The concept of virality and the structure of social media networks lend themselves to the use of SIR-like compartment models to study the spread of ideas between users on these platforms. This talk introduces the USBA – Unexposed, Sending, Bored, and Acclimated – family of discrete compartment models as a means of simulating how ideas reach and affect users and change the network structure, leading to the formation of echo chambers and polarization in sentiment.
• Author(s) : Adam Furman

[10988] THE IMPACT OF HABITAT LOSS ON A THREE-SPECIES TROPHIC SYSTEM

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Changes in ecosystems progress at a rapid pace mainly due to the climate crisis and human-induced perturbations. Researchers have used mathematical models to understand how species respond to these changes in habitat in order to ultimately forecast species extinctions and develop efficient conservation strategies. Our work highlights the fragility of predators hunting cooperatively under the loss of habitat.
• Author(s) : Jorge Duarte, Cristina Januario, Nuno Martins

[11023] A novel conservative Allen-Cahn system with structure-preserving property

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : A novel conservative Allen-Cahn (CAC) equation is presented in this poster. In existence, CAC equations have motion by mean curvature. Using the curvature-dependent Lagrange multiplier, proposed CAC equation has structure-preserving property. The structure-preserving property of the proposed CAC equation is well demonstrated through various numerical computational experiments.
• Author(s) : Soobin Kwak, Junseok Kim

[11036] Modeling and simulation of mini-grids under uncertainty

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Mini-grids generate and distribute energy locally and offer a reliable solution to ensure access to energy in countries where electrification is slow. Green energy generation, demand and weather naturally introduce uncertainties whereas the installation of a battery energy storage system asks for consideration of battery degradation. As thermal issues can significantly affect battery lifetime, an optimal control problem for the daily operation including thermal battery management under uncertainties is set up and solved numerically.
• Author(s) : René Henrion, Dietmar Hömberg, Nina Kliche

[11119] HEAT TRANSFER ENHANCEMENT IN SODIUM ALGINATE BASED MAGNETIC AND NON-MAGNETIC NANOPARTICLES MIXTURE HYBRID NANOFLUID

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Sodium alginate (SA) based hybrid nanofluids are novel new generation of fluids for heat transfer. The thermo-physical properties of these fluids are very classic in comparison to common fluids. This study aims to examine mathematical the heat transfer enhancement in viscoplastic non-Newtonian based Cu-Fe3O4 hybrid nanofluid, flowing over a stretching/shrinking sheet. SA is being used as a non-Newtonian, viscoplastic base fluid with the addition of Cu and Fe3O4 as non-magnetic and magnetic nanoparticles.
• Author(s) : Abid Hussanan

[11199] Bayesian Parameter Estimation for Ambient Solar Wind Models

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The solar wind is an essential driver of space weather geomagnetic storms. A significant challenge in using first-principle solar wind models is estimating input parameters that can not be directly measured. Thus, we need to quantify the uncertainty of such input parameters on the solar wind. We perform global sensitivity analysis to understand which parameters influence the model output the most and learn the posterior distribution of the influential input parameters via Bayesian inference.
• Author(s) : Opal Issan, Boris Kramer, Enrico Camporeale

[11204] Analysis of the fractal dimension of multidimensional data: The case of local field potentials

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Multisite recordings of local field potentials (LFPs) provide insights into the information processing in the brain. The fractal dimension (FD) of LFPs can measure the complexity of information flows between neuronal ensembles. However, the local stationarity and noise limit the assessment of FD from biological data. We provide a method for accurately estimating the FD from raw LFPs and illustrate it on synthetic data and electrophysiological recording in the rat hippocampus.
• Author(s) : Julia Makarova, Ricardo Muñoz, Oscar Herreras, Valeri A. Makarov

[11207] A nonlinear mathematical model on the Covid-19 transmission pattern among diabetic and non-diabetic population

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We study a three compartment mathematical model describing the dynamics between Covid-19 infected, diabetic and non-diabetic populations. Basic properties such as non negativity and boundedness of solutions, existence and stability of equilibrium points are studied, along with derivation of basic reproduction number. We have discussed a novel technique to estimate key parameters, and using them we have performed some numerical experiments which validate our theoretical results.
• Author(s) : Monalisa Anand, P. Danumjaya, P. Raja Sekhara Rao

[11213] Clusterless Inference of Compression of Spatial Representation in Hippocampal Replay

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We present a hierarchical framework combining latent-distance-dependent hidden Markov models and maximum likelihood estimation methods to infer the compression of spatial memory representation between locomotion and sharp wave-ripple states from unsorted, population spiking activity in rat hippocampus. This framework can be used for linking representations between one behavioral state where the neural data is constrained by finite sampling and others where the neural data is sufficient for accurate unsupervised estimation methods.
• Author(s) : Xinyi Deng, Joshua Glaser, Loren Frank, Scott Linderman

[11219] The Spectra of the Randic matrix of graphs

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Let G be a simple finite graph with vertex set ${v_1,v_2,…v_n}$. Denote the degree of vertex $v_i$ by $d_i, 1 \leq i \leq n$. The Randic matrix(which has association with the Randic index of graphs(in chemical graph theory)) of $G$, denoted by $R(G) = [r_{i,j}]$, is the n*n matrix whose $(i,j)$-entry $r_{i,j}$ is $r_{i,j}=1/\sqrt{d_id_j}$ if $v_i$ and $v_j$ are adjacent in $G$ and 0 otherwise.We present results on spectra of the Randic Matrix of graphs.
• Author(s) : Devsi Bantva

[11233] Random batch Quasi-Ewald method for the simulations of charged particles under dielectric confinement

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We derive an analytic, fast convergent lattice summation formula for the interaction energy/force of dielectric confined charged particles based on a novel Quasi-Ewald splitting strategy, then further achieving O(N) scaling for N-particle simulations via random batch importance sampling in the reciprocal space. The singularity in the analytical expression is carefully renormalized, thus extending our method to the case of metamaterials confinement, characterized by negative permittivity values.
• Author(s) : Xuanzhao Gao, Zecheng Gan

[11240] On Existence of Approximate Solution for Nonlinear Volterra Random Integral Equation

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Here, we prove the existence of approximate results for a nonlinear Volterra-type random integral equation in separable Banach space under mixed generalized compactness, contraction and caratheodory conditions, and also the existence of the locally attractive solutions is proved under some certain monotonicity conditions
• Author(s) : SHETE SIDDHARTH GANESH

[11261] Shortest path problem for recruiting personnel

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Present the problem of an outsourcing in the state of Querétaro with personnel recruitment problems where the objective is to optimize resources: Recruit more personnel for automotive or plastic production plants with the fewest number of routes and little fuel.
  Present the solution using the dijkstra algorithm.
• Author(s) : Giovana Ortigoza Alvarez

[11265] A backward semi-Lagrangian method (BSLM) to solve nonlinear coupled KdV equations(NCKdV)

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : This work introduces a BSLM to solve NCKdV. The equations are represented as linear dispersive equations(LDE) along the trajectories of particles described by the nonlinear Cauchy problems(NCP) in the Lagrangian viewpoint. The proposed method employs the FDM and BDF2 to solve LDE. To solve NCP, a deferred correction method is applied. From numerical experiments, the errors obtained by the proposed method are more accurate, even with a larger time step than the compared Eulerian methods.
• Author(s) : Soi Ji, Soyoon Bak

[11266] An efficient algorithm for solving 1D coupled Burgers’ equations in a semi-Lagrangian framework

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In this work, we introduce an efficient algorithm for solving 1D coupled viscous Burgers’ equations. The main accomplishment of this work is to develop a stable high-order algorithm for the system of reaction–diffusion equations. In the proposed algorithm, an interpolation strategy for finding the remaining upstream points is designed to dramatically reduce the high computational cost of solving the nonlinear Cauchy problem without damage to the order of accuracy.
• Author(s) : Jiseong Hur, Soyoon Bak

[11287] A population dynamics model for the information spread under the effect of social response

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We construct and analyze a mathematically reasonable and simplest discrete time one dimensional population dynamics model based on Mark Granovetter’s idea for the spread of a matter (rumor, innovation, etc.) in a population. Individual threshold values with respect to the decision making on the acceptance of a spreading matter are distributed throughout the population. We give the mathematical results on how the equilibrium acceptor frequency depends on the nature of threshold distribution in the population.
• Author(s) : Hiromi Seno, Reina Uchioke, Emmanuel J. Dansu

[11323] Opportunistically Stochastic Shortest Path Problems: From PDEs to AV-Routing

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We consider a class of Opportunistically Stochastic Shortest Path Problems (OSSPs) on graphs in which the decision maker chooses either a stochastic or deterministic transition to successor nodes. We provide a formal definition of OSSPs and develop conditions under which label-setting methods (e.g., Dijkstra’s or Dial’s Method) are applicable. We apply our results to examples of OSSPs arising in discretizations of anisotropic Hamilton-Jacobi PDEs and autonomous vehicle (AV) routing on lane-level road networks.
• Author(s) : Mallory Gaspard, Alexander Vladimirsky

[11352] Persistent homological figure detection technology and the latest status of its applications

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We first introduce persistent homological figure detection technology (developed by the presenter in 2022-23 (in press)). We will see that this technology detects figures using the death points of persistent barcodes.
  Then we explain how to apply this technology to several image analysis problems. We also discuss our current trial to use this technology for non-image data analysis and our future vision for a broader range of applications.
• Author(s) : Haruhisa Oda

[11362] Boundary layer preconditioners for elliptic problems in two dimensions

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We provide preconditioners for the linear system of equations that results from the discretization of second-order elliptic boundary layer problems using the least squares spectral element method. These preconditioners are constructed using the separation of variable techniques and being diagonalizable, they are simple to invert. Numerical results confirm the efficiency and robustness of the preconditioners.
• Author(s) : Aliya N. Kazmi, Akhlaq Husain, Ziya Uddin

[11391] Industrial Problem Solving Workshop Mexico. 16th years improving math collaboration between companies and academia.

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In this poster I will present a Mexican initiative that started 16 years ago in the Center of Research in Mathematics (CIMAT) to improve the collaboration between companies and academia in applied mathematics, statistics and computer science.
  Also, successful companies stories and new business models developed in this years, and the interaction with similar initiatives in other countries including the Artificial Intelligence Alliance in Mexico
• Author(s) : Sanchez-Bravo Ivete

[11396] Modelling transmission dynamics of Lassa fever transmission with two environmental pathway transmissions

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Lassa Fever (LF) is an animal-borne disease endemic in Africa, whose contaminated environment plays a vital role in its transmission dynamics. This study used a deterministic model to examine LF transmission with two environmental pathway transmissions. First, the stability of the model is established in regards to the model’s basic reproduction number, ${R_0}$. Finally, the model implements the sensitivity analysis to identify the parameters that fuel the LF spread using the Latin hypercube (LHS).
• Author(s) : Chinwendu E. Madubueze

[11404] Anomalous diffusion and chaotic motion in coupled standard map lattices

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The coupled SM processes exhibiting strong chaos and anomalous diffusion have been extensively studied. Although individual maps contain accelerator modes that cause anomalous transport, we observed that the coupled system’s global diffusion behavior is determined by the specific configuration of the imposed coupling. By estimating the average-diffusion properties for ensembles and measuring the strength of chaos, we find conditions and system arrangements that favor the suppression of anomalous transport and long-term convergence to normal diffusion.
• Author(s) : Henok Tenaw Moges, Thanos Manos, and Charalampos Skokos

[11422] Modeling the effect of unemployment and mass media on illicit drug use and terrorism dynamics

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The acts of illicit drug use and terrorism have negatively affected the economy and development of some nations because the death rate due to these two phenomena is increasing daily, particularly
  among the young generation. To this effect, a mathematical model was proposed, and cost-
  effectiveness analysis was conducted to ascertain the most effective and least expensive strategy required
  for preventing and controlling the burden of illicit drug use and terrorism in a population.
• Author(s) : John Olajide Akanni

[11425] Study on decoupled projection method for Cahn-Hilliard equation

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We study the numerical analysis for the Cahn–Hilliard (CH) equation using the decoupled projection (DP) method. Many kind of numerical schemes have been proposed to solve the CH equation. We consider the DP method for some schemes to verify the relation for the CH equation. We present the numerical experiments to demonstrate our analysis. As a future work, we will construct a novel numerical scheme using the relation with existing numerical schemes.
• Author(s) : Gyeonggyu, Lee

[11431] Estimation of Parameter Distributions for Reaction-Diffusion Equations with Competition using Aggregate Spatiotemporal Data

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : It is commonly assumed that individuals in a population have homogeneous diffusion and growth rates. This assumption can be inaccurate when the population has many subpopulations. We introduce the random differential equation version of the reaction-diffusion model with competition. We then rely on the Prokhorov metric framework to estimate joint distributions of diffusion and growth rates. We find that the random differential equation is more capable at predicting the cell density compared to other models.
• Author(s) : Kyle Nguyen, Erica M. Rutter, Kevin Flores

[11440] Multigrid solver with super-resolved interpolation

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The multigrid algorithm is an efficient numerical method for solving elliptic partial differential equations. The prolongation operator within the multigrid algorithm lends itself to a data-driven treatment with deep learning super resolution. We (i) propose the integration of a super resolution generative adversarial network (GAN) as the prolongation operator and (ii) explore the convergence properties of this super resolution-aided multigrid method on a class of multiscale PDEs typically solved in physics and engineering simulations.
• Author(s) : Francisco Holguin, GS Sidharth, Gavin Portwood

[11523] Boros integral associated with generalized Galué type Struve function

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Aim of this paper is to investigate Boros integral with three parameters involving generalized Galué type Struve function. Outcome of results are expressed in terms of the generalized Wright hypergeometric function. Several interesting corollaries of various struve functions are deduced as special case. Applications of obtained results are useful in applied mathematical sciences.
• Author(s) : Dr. Naresh Menaria

[11527] Gathering a robot swarm using circulant communication strategies

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We investigate the behavior of a swarm of autonomous, mobile robots moving in the plane. Robots observe positions of their neighbors according to a circulant interaction network and adapt their movement in order to gather in a single point. Using methods from continuous dynamical systems theory, we uncover conditions on the adaptation laws to guarantee gathering will be reached as well as a hierarchy of initial configurations of robot’s positions in terms of convergence speed.
• Author(s) : Jannik Castenow, Michael Dellnitz, Raphael Gerlach, Sören von der Gracht, Jonas Harbig, Friedhelm Meyer auf der Heide

[11540] Convergence Results based on Graph-Reich Contraction in Fuzzy Metric Spaces with Application

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : This article introduces a novel class of Reich-type contractions that meet the graph preservation criteria in the context of complete fuzzy metric spaces. Our key result is the natural extension of fuzzy metric spaces enriched with a graph, which adds the understanding of fixed points in metric spaces within the realm of graph structure. The findings are further supported by examples and applications.
• Author(s) : Shamoona Jabeen, Mehmet Emir Koksal, Mudasir Younis

[11553] Mood Prediction for Bipolar Disorder Patient with Sleep Pattern Information

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Mood prediction is an essential task for the treatment of bipolar disorder patients. Various
  mood prediction models were developed for bipolar disorder patients. However, the mood
  prediction model with probability threshold using sleep pattern information is not clearly
  developed. We propose the use of a Markov chain to predict the mood. Furthermore, we
  investigate the predictability difference between with and without the sleep pattern
  information.
• Author(s) : Dongju Lim, Yun Min Song, Jae Kyoung Kim

[11587] PC-SwinMorph: Patch Representation for Unsupervised Medical Image Registration and Segmentation

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Medical image registrations are fundamental tasks for different clinical procedures, where most solutions are supervised techniques. However, those techniques rely on well-representative datasets with ground truth, which is not always possible. To address this challenge, we propose a novel unsupervised framework for image registration. We first propose a patch-based contrastive strategy that enforces local feature representation. Secondly, we propose a patch-stitching strategy to eliminate artifacts. We demonstrate that our technique outperforms current state-of-the-art unsupervised.
• Author(s) : Lihao Liu, Zhening Huang, Pietro Liò, Carola-Bibiane Schönlieb, Angelica I. Aviles-Rivero

[11588] Far from equilibrium non-conserving exclusion process with site-wise dynamic defects

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : This study investigates a TASEP with non-conserving dynamics where the inhomogeneities on the lattice appear in the form of defects that stochastically bind and unbind the lattice. The influence of defects at the boundary sites is considered, due to which the entry rate of particles is affected. Utilizing mean-field approximations, we characterize the stationary state properties of the system and investigate the evolution of the phase diagram with respect to binding constant and obstruction factor.
• Author(s) : Nikhil Bhatia, Arvind Kumar Gupta

[11594] High-Efficiency 3D Video Coding Based On Machine Learning

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : With the expeditious growth of 3-dimensional (3D) video services for industrial applications, the accompanying large amount of video information requires innovative and powerful compression technologies to deliver more efficient video data compression. In this poster, we will exploit an amalgamation of artificial intelligence and machine learning together with data mining technology, linear regression scheme, and convex optimization method to realize a real-time intelligent low-power and high-efficiency 3D texture-and-depth coding algorithm modeling and hardware design.
• Author(s) : Jui-Hung Hsieh, Kuan-Yi Kuo, and Wei-Ting Chen

[11643] Medical Judgment Assistant: Data Classification base on Mahalanobis Distance

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : In this study, we are going to predict the medical data. We use the Mahalanobis distance to make an effective similarity on the data. The Mahalanobis distance is a distance measure based on the correlation between variables which can distinguish different types of overlapping data in the solution space, thereby improving the accuracy of data prediction and effectively helps doctors to make medical judgments.
• Author(s) : Kun-Huang Chen, Ming-Hsuan Chen, Wei-Jie Liang

[11649] IPPL 2.0: A massively parallel performance portable C++ Particle-in-Cell framework

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We present a performance portable C++ framework for Particle-in-Cell (PIC) methods, known as IPPL (Independent Parallel Particle Layer). IPPL makes use of Kokkos and HeFFTe (both part of the Exascale Computing Project), and MPI (Message Passing Interface) to obtain a massively parallel performance portable code which works across various hardware architectures. We showcase the performance of the latest IPPL version using examples from charged particle dynamics on state-of-the-art high-performance computing resources.
• Author(s) : Sonali Mayani, Antoine Cerfon, Tobia Claglüna, Matthias Frey, Severin Klapproth, Michael Ligotino, Veronica Montanaro, Sriramkrishnan Muralikrishnan, Alessandro Vinciguerra, Andreas Adelmann

[11672] Optimized first order alternating algorithms for fast and accurate low rank tensor decomposition

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : CP tensor decomposition has been proven to be a powerful tool for extracting information from large high order tensor, being widely applied in many areas such as chemistry, biology and medical science. However, efficiently computing the CP tensor still remains a challenge. In this study, we propose some optimized first order alternating least square algorithms for low rank tensor decomposition. We validate and illustrate the proposed algorithms by using simulated and real multi-way data.
• Author(s) : HUIWEN YU, OVE CHRISTIANSEN

[11673] AI-based numerical method to solve 2-dimensional fluid flow problem

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : This paper presents the numerical solution of 2-dimensional fluid flow problem using physics informed neural networks. The applicability of different wavelets as activation functions is investigated. As PINN depends on various parameters, the impact of network architecture on the accuracy of the model is also studied. The wavelet activation function represents an alternative to the tanh activation function and works better depending on the problem. It has been observed that the proposed method improves accuracy.
• Author(s) : Sai Ganga, Ziya Uddin, Rishi Asthana

[11758] Localized and spreading chaos in nonlinear multidimensional disordered lattices

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We implement the Generalized Alignment Index (GALI) method of chaos detection to investigate the dynamical behavior of nonlinear disordered lattice chains in one spatial dimension. We determine the probability to observe chaotic behavior as the system is approaches its linear limit. We also discriminate between localized and spreading chaos, with the latter dominating the dynamics for higher energy values.
• Author(s) : Bob Senyange

[11764] Representation Learning for Continuous Single-cell Biology with Graph Neural Networks

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Single-cell RNA sequencing provides high-resolution transcriptomics to study cellular dynamic processes, yet its high-dimensionality, sparsity, and noises undermine the performance of downstream analysis. We propose a deep learning framework based on Variational Graph AutoEncoder to learn a low-dimensional representation that preserves global information and local continuity. By applying pseudotemporal ordering to the extracted features, we show that the model accurately preserves the dynamic cell trajectories of real and synthetic scRNA-seq datasets.
• Author(s) : Chengkai Yang

[11773] Why Deep Surgical Models Fail?: Revisiting Surgical Action Triplet Recognition through the Lens of Robustness

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Surgical action triplet recognition is of high relevance as it provides the surgeon with context-aware support and safety. The go-to strategy develops new network mechanisms. However, the performance of state-of-the-art techniques is substantially lower than other surgical tasks. Why is this happening? This is the question that we address in this work. We present the first study to understand the failure of existing deep learning models through the lens of robustness and explainability.
• Author(s) : Yanqi Cheng, Lihao Liu, Shujun Wang, Yueming Jin, Carola-Bibiane Schönlieb, Angelica I. Aviles-Rivero

[11826] Analysis of viscous dissipative flow of Casson hybrid nanofluid at the stagnation point over a rotating sphere

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The flow of non-Newtonian Casson hybrid nanofluid near the stagnation zone over the rotating sphere considered. The influence of thermal radiation, magnetic field and viscous dissipative effects are also taken into account. Similarity transformations are used to convert the governing partial differential equations into ordinary differential equations. The velocity and temperature profiles are graphically displayed for different parameters involved in the study. Nusselt number and skin friction coefficient are also computed.
• Author(s) : Tanvi Singla, Sapna Sharma, Bhuvaneshvar Kumar

[11827] The influence of road capacity on traffic flow in a percolation-backbone fractal with onramp

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : To study the influence of variable road capacity and onramp effect on traffic flow dynamics in a network, percolation backbone fractal network is considered. The fractal network is described with the help of cell transmission graph. By using the speed matching model, the density equations are derived. The urban scale macroscopic fundamental diagrams are obtained numerically in the fractal network. It is observed that as the capacity of road decreases, the traffic flow decreases.
• Author(s) : Muskan Verma, Sapna Sharma

[11830] Effect of adding reactions on the chemical reaction network sensitivity

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Biological functions arise from the intricate dynamics of reaction networks comprising numerous reactions and chemicals. However, network information is often inaccurate and diverse across species. Previously, we developed the “Structural Sensitivity Analysis”, which enables the responses of reaction systems to parameter perturbations to be determined solely from network topology. In this study, we investigate how small alterations to network structure affect system behavior. The results can be classified into five distinct cases based on topology.
• Author(s) : Atsuki Hishida, Atsushi Mochizuki

[11839] A generalized structural bifurcation analysis of chemical reaction networks

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Chemical reactions link metabolites and form complex networks in living cells. We have previously developed “structural bifurcation analysis,” by which bifurcation properties of reaction systems are determined solely from network topologies. In this work, we establish a precise formalization connecting our analysis to conventional methods based on Jacobian matrices. The formalization increases applicability of the analysis, e.g. determining multistationarity, without assuming the full-rankedness of stoichiometric matrices or eliminations of equations/chemicals.
• Author(s) : Yong-Jin Huang, Takashi Okada, Atsushi Mochizuki

[11855] Quadratic Lie algebras algorithms applied over oscillator algebras

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Quadratic Lie algebras appear in Mathematics and Physics. Main examples are oscillator and generalized oscillator which are related to space-time models and determine some Lie groups with Lorentz metrics or Lorentzian cones. This variety of algebras with bilinear invariant forms can be built using double extensions from a metric vector space via derivations. In this poster we will see an overview of how all these concepts can be algorithmically obtained. Available in our Github repository.
• Author(s) : Jorge Roldán-López, Pilar Benito

[11875] Interplay of reservoirs in an exclusion process with limited resources

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We study a conserved system comprised of two directed identical lanes connected to two distinct reservoirs having finite resources. Our findings display two distinct phases that admit shocks at steady-state. One phase admits a delocalized shock in each lane with perfectly synchrony while the other phase, the single shock in the system may traverse both lanes or remain restricted to a single lane, depending upon the size of the system.
• Author(s) : Arvind Kumar Gupta, Bipasha Pal

[11901] Convergence rates of consensus in multi-agent systems with communication delays

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In the present talk, we discuss convergence rates of consensus in multi-agent systems with two types of communication delays: discrete and continuous. The dynamics of the agents is described by functional differential equations. We carry out the numerical simulations on some networks involving the delays. Then, by using a Lyapunov functional, we estimate the convergence rates in each case. The convergence of the continuous delay system is faster than that of the discrete one.
• Author(s) : Yoshinori Katanaya,Rijyo Yamakawa,Hirokazu Komatsu,Hiroshi Yokota

[11943] A MATHEMATICAL MODEL FOR DECIPHERING THE IMPACT OF OBESITY ON CANCER

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : This work provides insight into how obesity contributes to cancer progression. For this, we developed a diffusive obesity-cancer model consisting of cancer cells, normal cells, fat cells, macrophages, and an ECM. We have directed the formed model’s global existence and non-negativity. We presented a traveling wave analysis and calculated the minimum wave speed. Numerical simulation (in both 1D and 2D) discloses that cancer spread increases with increased haptotaxis coefficient and growth rate of obese cells.
• Author(s) : Ani Jain, Parimita Roy

[11965] The Influence of Human Behavior in COVID-19 Modeling

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We incorporate human behavior in a disease model to study how it can affect the spread of COVID-19 in small college settings over short periods of time. We obtain dampened oscillations by introducing a risk assessment function, which allows us to account for participants’ reactions to the spread of the disease and institution policies. We compare these oscillations to COVID-19 data from US postsecondary institutions and discuss the role risk assessment plays in COVID-19 management.
• Author(s) : Ognyan Simeonov

[11995] Asymptotic tracking of a point cloud moving on Riemannian manifolds

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We present two Cucker-Smale type models for the asymptotic tracking of a point cloud moving on complete, connected, and smooth Riemannian manifolds. For each model, we provide a sufficient framework in terms of a moving target point cloud, system parameters, and initial data. In the proposed framework, we show asymptotic flocking, collision avoidance, and asymptotic tracking to a given point cloud. The main result is a joint work with Hyunjin Ahn, Seung-Yeal Ha and Jaeyoung Yoon.
• Author(s) : HYUNJIN AHN,SEUNG-YEAL HA,JAEYOUNG YOON

[12006] Perovskites oxides and their theoretical modelling

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Transition metal atom doping in Perovskite material is accomplished to investigate the structural, chemical and magnetic properties of synthesized samples. Single phase, cubic crystal formations of BasNO3 are revealed from the structural properties. Transmission electron micrographs (TEM) display the formation of polygonal discs with nanoscale (~50 nm) dimensions. Density functional theory (DFT) is employed to optimize the structural para￾meters obtained from XRD analysis. A (3 × 2 × 2) supercell of BaSnO3 was created for
• Author(s) : Ishtihadah islam, Seemin rubab

[12015] Application of machine learning to predict dynamics of epidemiological models that incorporate human behavior

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In this work, we present modeling, analysis and simulation of a mathematical epidemiological model which incorporates human social, behavioral, and economic interactions. We discuss an approach based in Physics-Informed Neural Network, which is capable of predicting the dynamics of a disease described by modified compartmental models that include parameters, and variables associated with the governing differential equations. Finally, human behavior is modeled stochastically and it is included in the compartmental models.
• Author(s) : Alonso Ogueda-Oliva, Padmanabhan Seshaiyer

[12047] Impacts of uncertainty about historical information on downstream

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In traffic environment, complex uncertainties such as network fluctuations, driver personality, and traffic disruption may affect traffic information. Therefore, by considering uncertainty about historical density information (UHDI), a new lattice model is developed. The UHDI effect is probed using linear and nonlinear stability analysis. The instability is found with an increased value of the UHDI coefficient. The modified Korteweg-de-Vries equation is obtained to describe congestion’s characteristics. Finally, numerical simulations are implemented to confirm theoretical findings.
• Author(s) : Ms.Daljeet Kaur, Dr.Sapna Sharma

[12065] Variational Approach to Hamiltonian Random Impulsive Differential Systems

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We establish a variational framework for Hamiltonian differential systems with random impulses. Using the generalized saddle point theorem, we show that the related energy functionals have multiple critical points, that is, the systems have multiple weak solutions. Moreover, the sufficient conditions for multiple solutions are investigated. We finally give an example to illustrate the feasibility and effectiveness of this method. The result has been published on Qualitative Theory of Dynamical Systems.
• Author(s) : Dan WU

[12086] CPFloat: A C Library for Simulating Low-Precision Arithmetic

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : CPFloat is a full mathematical library, written in C, for simulating low-precision arithmetic. CPFloat exploits the bit-level floating-point representation of the format in which the numbers are stored, by relying only on low-level bit manipulations and integer arithmetic. In numerical experiments, the new techniques bring a considerable speedup (typically one order of magnitude or more) over existing alternatives in C, C++, and MATLAB.
• Author(s) : Massimiliano Fasi, Mantas Mikaitis

[12200] Diatom Identification in Microscopy Videos Using Computer Vision Techniques

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We present a novel approach for identifying diatoms in microscopy videos using convolutional neural networks (CNNs). Our method takes advantage of the CNN’s ability to learn features that are important for identifying diatoms, such as their unique shape and texture. We demonstrate the effectiveness of our approach on a dataset of microscopy videos of diatoms. Our work has important implications for the field of diatom analysis, where accurate identification
• Author(s) : Anton Phipps

[12227] Deciphering tissue-specific gene-drug patterns by sparse tensor partial least squares method

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Exploring the underlying relationship between drugs and genes provides novel insights into molecular mechanisms and potential drug targets for therapy. In order to decipher the gene-drug association patterns shared by multi-type tissues or across the specific tissue, we developed a sparse tensor partial least squares method to simultineously integrate the large-scale pairwise gene expression and drug response data sets, deriving from diverse tissue types. We identified 10 comodules with significantly coordinated gene-drug associatons.
• Author(s) : Chen J, Min W

[12248] On Q-integral graphs with Q-spectral radius 6

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : A connected Q-integral graph with Q-spectral radius six is either known has maximum edge-degree five or is a bipartite graph containing a specific induced subgraph. In this talk, we will improve this result by showing that there is no connected Q-integral bipartite graph with Q-spectral radius six by checking numerous matrix eigenvalues.
• Author(s) : Semin Oh, Jeong Rye Park, Jongyook Park, Yoshio Sano, Jeongmin Ha, Sangkon Han

[12256] Modeling and Numerical Simulation of Two-Dimensional TMDC Memristive Devices

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In recent years, 2D layered transition metal dichalcogenides (TMDCs) received a great deal of attention as promising memristive materials for neuromorphic computing applications. Despite extensive experimental progress, the current work on lateral TMDCs lacks a deep physical understanding. We present a mathematical model and solve it numerically. By including an ionic species and Schottky barrier lowering in a self consistent fashion, we can explain the hysteresis in current-voltage curves.
• Author(s) : Benjamin Spetzler, Dilara Abdel, Patricio Farrell

[12277] Stocastic model for location and dispatching ambulances with partial coverture

• Date & Time : P (Aug., 12:20-13:20)
• Abstract : A stochastic model for ambulance location and dispatching with a partial coverage Resumen: Ambulance location and dispatch problems have been used to improve emergency medical service systems for patients’ attention. Commonly ambulances available in these systems are not enough to cover all emergency calls. We consider partial coverage for attending emergency calls to raise the number of demand points covered by at least one ambulance. Considering this partial coverage, we present an integer stochastic problem.
• Author(s) : Beatriz A. García-Ramos , Yasmín Á. Ríos-Solís

[12297] Modeling earthquake process and ground motion based on a stochastic differential equation

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The faulting processes of earthquakes exhibit considerable variation and can be modeled using a stochastic process framework. To that end, we have developed a model that represents the time series of fault slip as a convolution of two solutions of the Bessel stochastic differential equation. We have evaluated the validity of the model in light of empirical laws on the faulting processes and demonstrate its application to ground motion prediction.
• Author(s) : Shiro Hirano,

[12300] Features Engineering and Machine Learning Methods for the Prediction of the Patients’ Postoperative WOMAC Score After Total Knee Replacement

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The purpose of this paper aims to study the factors and creates a model to predict the patients’ postoperative WOMAC score after total knee replacement. The influencing factors are found by feature engineering using several techniques, e.g. Generalized Linear Model, Support Vector Machine, Deep Learning, and Gradient Boost Tree. Afterwards, the model was created by the Gradient Boost Tree technique which groups different attributes from feature engineering.
• Author(s) : Saranchai Sinlapasorn, Benjawan Rodjanadid , Jessada Tanthanuch, Bura Sindhupakorn

[12354] A Portable Platform for Label-free Virus Enrichment and Fast/Accurate Virus Surveillance using Raman Spectroscopy

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We presents a novel platform, named VIRRION (VIrus capture with Rapid Raman spectroscopy detection and IdentificatiON), which consists of a handheld microdevice designed to capture different viruses based on their sizes, and performance of optical identification using in-situ Raman spectroscopy coupled to a machine learning algorithm and database.
• Author(s) : Yin-Ting Yeh

[12356] Modeling CD4+ and CD8+ Cell Activity to Combination Immunotherapy in Mice with Triple Negative Breast Cancer

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The immune-tumor response aided by immunotherapy is analyzed through model optimization to our system of ordinary differential equations. Our equations describe the core biology of an immunogenic anti-tumoral response from CD4+ and CD8+ cells which are observed by state-of-the-art longitudinal positron emission tomography (PET) images. Corresponding tumor volume measurements are also included with the data. Immunotherapeutic effects on our tumor-initiated inflammation model are validated by comparisons to immunotherapy single treatment data.
• Author(s) : Dayton Syme, Yun Lu, Anna G. Sorace, and Nicholas G. Cogan

[12367] A modified approach for fractional transportation problem under interval-valued Fermatean fuzzy sets

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Akram et al. (AIMS Mathematics 7 (2022) 17327–17348) introduced a method to solve Fractional transportation problem (FTP), where parameters are represented by triangular interval-valued Fermatean Fuzzy Numbers (FFNs). Moreover, the optimality criterion was extended for transportation problem (TP) to FTP. But their approach is significantly more time consuming. To address this issue, we developed a ranking function to convert all FFNs to crisp numbers, and compared our results with Akram et al.’s method.
• Author(s) : Parul Tomar, Amit Kumar

[12371] Inappropriateness in simple non-cooperative games with intuitionistic fuzzy information

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Yang et al. (Appl. Intell (2021) 51: 6685-6697) proposed an approach to solve matrix games and bi-matrix games with intuitionistic fuzzy payoffs to reduce computational efforts. However, after a deep study of Yang et al.’s approaches, it is noted that a mathematically incorrect result is considered in their approach. Hence, it is inappropriate to use Yang et al.’s approach. In this note, the mathematically incorrect result, considered in Yang et al.’s approach, is pointed out.
• Author(s) : Kirti, Tina Verma, Amit Kumar

[12392] Exploring the Impact of Controlled Vehicles on Mixed Traffic in Cellular Automata

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : To investigate the impact of controlled vehicles (CV) on mixed traffic, we propose a controlled stochastic optimal speed (CSOV) model in cellular automata. The CSOV model extends the original SOV model by incorporating two different vehicle controls, the gap-based control and the flow-based control. We will discuss the simulation results under different penetration rates of the CV and strength of control parameter.
• Author(s) : Kayo Kinjo, Akiyasu Tomoeda

[12401] A 5th order finite difference WENO Scheme

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We propose a new Z-type nonlinear weights of the fifth-order finite difference WENO scheme for hyperbolic conservation laws. Specifically, we take the pth root of the smoothness indicators and follow the form of Z-type nonlinear weights, leading to fifth order accuracy in smooth regions, even at the critical points, and sharper approximations around the discontinuities. We also prove that the proposed nonlinear weights con- verge to the linear weights as p → ∞.
• Author(s) : Xinjuan Chen, Jiaxi Gu, Jae-hun Jung

[12412] Simulating Saudi Kidney Exchange Program

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Kidney Exchange matches one patient and his or her incompatible donor with another pair in the same situation for an exchange. Kidney exchange programs (KEPs) circumvent these barriers as they enable patients to exchange donors. Such a central KEP has the potential to match more patients and to find better matches, thus increasing the quality of life of transplant recipients and reducing mortality. We used simulation and modeling tools to evaluate novel matching strategies.
• Author(s) : Michal Mankowski, Khalid AlMeshari

[12419] Optimal strategy of non-pharmaceutical interventions considering medical capacity during the SARS-CoV-2 omicron-dominant period

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Given the relatively low severity and fatality rates of SARS-CoV-2 omicron variant compared to pre-omicron, strict non-pharmaceutical interventions (NPIs) that cause high economic costs may not be necessary. We investigated an optimal NPIs strategy considering medical capacity during the spread of omicron in the Republic of Korea using optimal control theory. Our study presents strategy with appropriate balance between returning to normal life and prohibiting the limited hospital beds for severe case from being overwhelmed.
• Author(s) : Yuna Lim, Youngsuk Ko, Renier G. Mendoza, Victoria May P. Mendoza, Jongmin Lee, Yubin Seo, Eunok Jung

[12425] Soft magnetic microrobots move more efficiently with a flat tire

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : This poster will describe the rolling of active Pickering emulsions – small droplets (~10-100 um) covered in smaller (~1um) active particles that can be rolled along a surface by an external, AC magnetic field. Curiously, these droplets roll much faster and more efficiently when they have a larger area of contact with the confining surface. The poster will outline experiments and numerical simulations that are used to quantify and explain this behavior.
• Author(s) : Brennan Sprinkle, Yan Gao, David Marr, Ning Wu

[12452] Numerical Study of Temperature Dependent Viscosity and Thermal Conductivity on a Natural Convection Flow over a Sphere in Presence of Magneto Hydrodynamics

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The objective is to investigate the numerical study of temperature-dependent viscosity and thermal conductivity on the natural convection flow of an electrically conducting fluid over an isothermal sphere in presence of magnetohydrodynamics. The governing equations are transformed into dimensionless non-similar equations by using a set of transformations and solved numerically. The computational findings for dimensionless velocity profiles, temperature profiles, local skin friction coefficient, and local heat transfer coefficient are displayed graphically and in tabular forms.
• Author(s) : Md. M. Alam, Rina Begum, Mohammad Mahfuzul Islam and M. M. Parvez

[12469] Computing solution space properties of combinatorial optimization problems via generic tensor networks

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We introduce a unified framework to compute the solution space properties of a broad class of combinatorial optimization problems. These properties include finding one of the optimum solutions, counting the number of solutions of a given size, and enumeration and sampling of solutions of a given size.
• Author(s) : Jin-Guo Liu, Xun Gao, Madelyn Cain, Mikhail D. Lukin, Sheng-Tao Wang

[12472] Robust train trajectory optimization

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Variating operating conditions may produce delays in railways. We model the Robust Train Trajectory Optimization (RTTO) problem aiming to minimize the impact of model parameter uncertainty on the calculated energy-efficient trajectories, which would be drivable under any of the considered operating conditions. We analyze RTTO using the Robust Maximum Principle, reformulate the problem as a Quadratically-Constrained Quadratic Programming problem and showcase the performance of the model in a real case study.
• Author(s) : Alex Cunillera, Ramon M. Lentink, Niels van Oort, Rob M.P. Goverde

[12502] Control problem for nonlinear fractional dispersive system

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Motivated by the porous media equation as well as population dynamics models, we study control problem for nonlinear fractional system of differential equations. The solving strategy that we propose is essentially novel and involve techniques and tools from fractional calculus and functional analysis. The basic idea consists of linearizing the system, and then using the fixed point results. Also, we are interested in concepts such as controllability and observability in the fractional framework.
• Author(s) : Maja Jolic, Sanja Konjik, Darko Mitrovic

[12503] How bifurcations of dynamical systems are helping train drivers to save energy in the Netherlands

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Nederlandse Spoorwegen is the main Dutch railway operator. Their drivers use an application that shows when to stop accelerating or braking to arrive timely while saving energy. Bifurcation theory is used to analyze the solutions of the dynamical system that models a train’s dynamics, to identify terminal speeds and to select the analytical solution of the model to be used in the calculations. The new advice algorithm currently in operation is based on this research.
• Author(s) : Alex Cunillera, Harm H. Jonker, Gerben M. Scheepmaker, Wilbert H. T. J. Bogers, Rob M. P. Goverde

[12516] Selection mechanism in non-Newtonian Saffman-Taylor fingers

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : An analytical method based on the WKB approximation is offered to forecast the finger width ($\Lambda$) of a Newtonian fluid pushing a non-Newtonian fluid by selecting a unique finger width from the family of possible solutions. We also discovered that the relationship between $\Lambda$ (dimensionless) and the parameters containing the viscosity and surface tension, $\nu$ (dimensionless), has the form: (I) $\Lambda \sim \frac{1}{2}-\mathcal{O}(\nu^{-1/2})$ (for shear-thinning) (II) $\Lambda \sim \frac{1}{2}+\mathcal{O}(\nu^{2/(4-n)})$ (for shear-thickening). Significant comparison is provided.
• Author(s) : D. Bansal, D. Ghosh, S. Sircar

[12556] Micro-Macro Modelling of Particulate Systems

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Mathematical Modelling of particulate systems is challenging because of various complex mechanisms acting on the system. The problems are two folds: (a) an accurate and efficient numerical method is required for solving the underlying system of integro partial differential equations, and (b) modelling of kinetics into the model parameters is extremely difficult. This work addresses both aspects of the problem.
• Author(s) : Jitendra Kumar

[12590] Drop Impact: modelling a lubrication air layer and surface waves in droplet rebound dynamics.

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : A small liquid droplet rebounding from the free surface of a deep bath has been studied experimentally and with a variety of models. An important part of the physics is a thin layer of air separating the droplet and the free surface. We develop a fully coupled dynamic model for the drop-air-bath interaction, using lubrication theory to deduce the pressure transfer between the drop and free surface in two-dimensions for both rigid and deformable impact.
• Author(s) : Kat Phillips, Paul Milewski

[12594] A monotone discretization for integral fractional Laplacian on bounded Lipschitz domains: Pointwise error estimates under Hölder regularity

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We propose a monotone discretization for the integral fractional Laplace equation on bounded Lipschitz domains with the homogeneous Dirichlet boundary condition. By using a discrete barrier function that reflects the distance to the boundary, we show optimal pointwise convergence rates in terms of the Hölder regularity of the data on both quasi-uniform and graded grids. Several numerical examples are provided to illustrate the sharpness of the theoretical results.
• Author(s) : Rubing Han, Shuonan Wu

[12600] OPTIMAL CELL AVERAGE DECOMPOSITION FOR HIGH-ORDER BOUND-PRESERVING SCHEMES

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The problem of seeking optimal cell average decomposition (OCAD) arises from constructing efficient high-order bound-preserving numerical methods within Zhang-Shu framework. It remained unknown what CAD is optimal for higher-degree polynomial spaces. We establish the general theory for analyzing the OCAD problem on Cartesian meshes in 1D and 2D, and rigorously prove that the classic CAD is optimal for general 1D $\mathbb{P}^k$ spaces and general 2D $\mathbb{Q}^k$ spaces, but not optimal for the 2D $\mathbb{P}^k$ spaces.
• Author(s) : Shumo Cui, Shengrong Ding, Kailiang Wu

[12606] An efficient optimization approach for three-dimensional packing problems

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : This study proposes a novel deterministic optimization approach for globally solving 3D packing problems that have been applied in various practical applications. The original problems are transformed into an equivalent linear mixed integer programming problem containing much less number of binary variables than current methods. Then an efficient algorithm is developed to solve the transformed problems. Experimental results reveal that the proposed method can effectively solve large scale 3D packing problems within a reasonable time.
• Author(s) : Jung-Fa Tsai, Ming-Hua Lin

[12651] Development of a mathematical model for adsorption of multiple components from a polluted fluid

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We develop a mathematical model to describe an adsorption process in which two species compete to occupy the available sites on the adsorbing particles. The governing equations are kinetic and advection-diffusion for each species, related both to the individual adsorption processes and describing the interaction between the two species. Once the relevant variables are scaled and the dimensionless parameters have been identified, the system is solved numerically and validated against experimental data.
  
• Author(s) : M. Calvo-Schwarzwalder, A. Cabrera-Codony, A. Valverde, M. Aguareles, T.G. Myers

[12662] Shaping up scientific Machine Learning

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The identification of interesting substructures within jets is an important tool for probing the Standard Model at colliders. We present SHAPER (Shape Hunting Algorithm
  using Parameterized Energy Reconstruction), a measure theoretic approach to identifying such substructures, by adapting modern advances from dictionary/manifold learning. Further, we show that SHAPER constitutes the optimal approach to such collections of problems in high energy physics and estimate optimization guarantees on the same.
• Author(s) : Demba Ba, Akshunna S. Dogra, Rikab Gambhir, Abiy Tasissa, Jesse Thaler

[12699] Fully automated scar quantification in myocardial infarction

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Cardiac magnetic resonance imaging is the current standard modality for assessing the state of the heart after myocardial infarction. Based on convolutional neural networks we present a new framework, which calculates the extent of myocardial infarction in a fully automated way. Results show very good agreement between manually and automatically calculated infarction volumes even outperforming state-of-the-art methods. Our algorithm could greatly reduce the time to diagnosis as well as support physicians in further treatment steps.
• Author(s) : Schwab Matthias, Pamminger Mathias, Kremser Christian, Haltmeier Markus, Mayr Agnes

[12722] Parameter Identification of Vegetation Pattern Dynamic Systems

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Vegetation patterns, like spots, and stripes, emerge in dryland regions. We propose a probabilistic method to identify parameters in vegetation pattern dynamic systems. Our approach determines the most probable parameter values that allow the model to most closely replicate observed vegetation patterns. By identifying parameters, the models can be validated and reveal key properties of the soil environment, like water redistribution rates. This could aid evaluation and prediction of vegetation in dryland regions.
• Author(s) : Xinyue Luo, Yu Chen

[12727] Dynamical Motion of Surface Active Flow Driven Droplets

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Active droplets which can autonomously locomote are of both fundamental interest and of practical importance. We use rigid multiblob method to simulate the dynamic trajectory of droplets near a no-slip wall mobilized by active point-like particles, which can induce hydrodynamic flows on its interface. We find that the simulated trajectories exhibit rich dynamical modes consisting of circular, helical, or petal-like motions, which can be controlled by the initial tilt angles and microswimmer configurations.
• Author(s) : Zheng Yang, Zecheng Gan, Rui Zhang

[12731] Strong cosmic censorship theorem in Bakry-Emery spacetimes

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : A class of naked strong curvature singularities is ruled out in Bakry-Emery spacetimes by using techniques of differential topology in Lorentzian manifolds. These spacetimes admit a Bakry-Emery-Ricci tensor which is a generalization of the Ricci tensor. This result supports to validity of Penrose’s strong cosmic censorship conjecture in scalar-tensor gravitational theories, which include dilaton gravity and Brans-Dicke theory.
  
• Author(s) : Makoto Narita

[12751] AN ANALYSIS OF THE INTERACTION EFFECTS OF SOCIOECONOMIC AND DEMOGRAPHIC FACTORS ON COVID-19 IN THE UNITED STATES DURING PRE-VACCINATION PERIOD: EMPIRICAL EVIDENCE FROM NEGATIVE-BINOMIAL REGRESSION MODELS

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In this study we analyze the variability in infection and fatality rates due to COVID-19 across the US during the pre vaccination period (January 2020 – December 2020). For a better understanding, the pre vaccination period is divided into two phases based on the sharp rise in COVID-19 deaths and each phase clustered based on infection and fatality rate.We studied the main and interaction effects of several risk factors using Negative Binomial regression modeling.
• Author(s) : Sucharitha Dodamgodage, Dinushani Senarathna, Stephanie Andreescu, , James Greene, Shantanu Sur, Sumona Mondal

[12791] Representation and use of finite elements in Firedrake.

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The finite element method is based on finding approximate weak solutions to variational problems. These solutions live in finite element spaces.The way we represent and encapsulate the definition of the elements making up the spaces is key to how the software is eventually written and used in the future, as well as its extensibility.This work will look at how this is done in Firedrake and how it might be improved.
• Author(s) : India Marsden, David A. Ham, Patrick E. Farrell

[12812] Solution of split inverse problems using fixed point iterations

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : This study proposes a new inertial Mann-type Tseng’s extragradient algorithm to approximate the solution of split inverse problems in real Hilbert spaces. We establish strong convergence of the proposed scheme to a minimum norm for monotone and uniformly continuous single-valued operators with self-adaptive step size, provide numerical implementations to illustrate the convergence of our method, and compare it with the non-inertial version and existing related algorithms.
• Author(s) : Olaoluwa Ogunleye, Timilehin O. Alakoya, Oluwatosin T. Mewomo, Olaniyi S. Iyiola

[12813] Designing good teaching materials to train future applied mathematicians

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : This poster presents a bundle of activities designed for differential equations instructors to foster cultural competence and mathematical skills in students. We use frameworks of cultural responsiveness to create a diverse and inclusive classroom environment. We introduce the idea of academic agreement and offer best practices for teachers to design similar materials on different topics for different courses, preparing future applied mathematicians for modeling work. This work is submitted to a math education journal.
• Author(s) : Yanping Ma, Gail Tang

[12828] Partial impulse observability of linear descriptor systems

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : A research paper in Systems & Control Letters, vol. 61, no. 3, pp. 427–434, 2012, by M. Darouach, provides a functional observer design for linear descriptor systems under the partial impulse observability condition. The observer design is correct, but there was a flaw in the algebraic criterion characterizing partial impulse observability. In this poster, we derive a novel characterization of partial impulse observability in terms of a simple rank condition involving the system coefficient matrices.
• Author(s) : Juhi Jaiswal, Thomas Berger, Nutan Kumar Tomar

[12858] Inertia-Based Natural Frequency Re-Assignment of a Real Reciprocating Hydrogen Compressor Used in Refineries

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Torsional resonance problem of the reciprocating hydrogen compressor used in refinery factory can cause fire and explosion hazards. The events result in manufacturing interruption and hence immense economic loss. Although such problem can be mitigated if shifting or re-assigning the driver-compressor system’s natural frequencies (TNFs) can be achieved, the desired natural frequencies are usually difficult to realize because of the uncertainties involved in the system’s mathematical modelling. In this paper, a torsional resonance problem occurred
• Author(s) : Jin-Wei Liang

[12874] The effect of directional dispersal of predator on predator-prey model

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In this talk, we present the effect of directional dispersal of a predator on a predator–prey model. The prey is assumed to have traits making it undetectable to the predator and difficult to chase the prey directly. Directional dispersal of the predator is described when the predator has learned the high hunting efficiency in certain areas, thereby dispersing toward these areas instead of directly chasing the prey. We investigate the stability of the semi-trivial solution
• Author(s) : Kwangjoong Kim, Wonhyung Choi, Youngseok Chang and Inkyung Ahn

[12907] On fluid–structure interactions with the Coulomb friction law boundary condition

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We propose a model in a fluid–structure interaction system composed by a solid and a viscous incompressible fluid, using Coulomb friction law. The fluid can slip on the boundary if the tangential component of the stress tensor is large. In the opposite case, we recover the Dirichlet condition. The governing equations are the Navier–Stokes system for the fluid and Newton laws for the body. We prove there exists a weak solution and some numerical results.
• Author(s) : Loredana Balilescu, Jorge San Martin, Takeo Takahashi

[12930] Multiscale computations for the elastic quadratic eigenvalue problem in composite structure

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : A multiscale asymptotic analysis and computational method based on the Second-Order Two-Scale(SOTS) approach is proposed for the elastic quadratic eigenvalue problems in the periodic composite domain. Both the second-order expansions of the eigenfunctions and eigenvalues are derived and the corresponding finite element procedures are established. Linearized methods are introduced to solve the homogenized quadratic eigenvalue problems and the effectiveness of the second-order asymptotic model is demonstrated.
• Author(s) : Periodic domain; Asymptotic homogenization; Quadratic eigenvalue problem; Finite element algorithm

[12944] Optimization-based approach for computing equilibrium shapes of crystals

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Computing equilibrium shapes of crystals (ESC) is an important problem in materials science. It can be described as minimizing anisotropic energy functional with a prescribed mass constraint. The highly nonlinear, singular anisotropic terms make this problem very challenging at both the analytical and numerical levels. Especially for strongly anisotropic cases, the ESC will form some singular, sharp corners. We proposed an optimization-based approach to predict the ESC and gave its convergence results.
• Author(s) : Zeyu Zhou, Wen Huang, Wei Jiang, Zhen Zhang

[12970] Effect of population distribution on critical time of reaction-diffusion systems

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : A finite measure of the time to approach a steady state, known as critical time, is important in many practical situations. Here, we consider reaction-diffusion processes with certain types of boundary conditions which can produce different kinds of asymptotic spatially homogeneous steady-states. We investigate numerically how the initial distributions of populations can affect the critical time. In this study, we started with a single reaction-diffusion equation before moving to reaction-diffusion systems.
• Author(s) : Nor Farah Wahidah Binti Nor Khalid, Mohd Almie Bin Alias

[12974] Mittag-Leffler stability for a fractional Klein–Gordon equation

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We consider the time fractional Klein-Gordon equation, which is a relativistic version of the well known Schrödinger equation. The existence and uniqueness are asserted using standard results from the theory of fractional resolvents. The Mittag-leffler stability is established by cosntructing suitable energy functionals. Apart from the nonexistence of semigroup solutions, the integer order chain rule is no longer valid and we overcome this difficulty by making use of a modified product rule for fractional case.
• Author(s) : Kausika Chellamuthu

[13023] Nonlinear Model Reduction for Slow-Fast Stochastic Systems near Unknown Invariant Manifolds

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We introduce a nonlinear model reduction technique for high-dimensional stochastic systems having a low-dimensional invariant effective manifold with slow dynamics, and high-dimensional, large fast modes. Given only access to short bursts of simulation, we design an on-the-fly algorithm exploring the effective state space efficiently, without losing consistency with underlying dynamics. This construction enables fast, efficient simulation of the effective dynamics that averages out fast modes, plus estimation of crucial features and observables of such dynamics.
• Author(s) : Sichen Yang, Felix X-F Ye, Mauro Maggioni

[13038] Stability of an inverse problem for Biot’s consolidation system in poro-elasticity

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We consider a coupled system of mixed hyperbolic-parabolic type which describes Biot’s
  consolidation model in poro-elasticity. We study an inverse problem of determining five spatially varying coefficients in the model of Biot’s consolidation system by three measurements of displacement in an arbitrary subboundary and temperature in an arbitrary neighborhood of the boundary over a time interval. We prove a logarithmic stability estimate for the inverse
  problem.
• Author(s) : Wensheng Zhang, Zifan Jiang

[13043] Competitive Adsorption Processes Applied to Contaminant Removal

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We introduce the basic model describing the roles of diffusion, advection and adsorption during the removal of one single contaminant from a carrier fluid and extend it to account for competitive adsorption, a phenomenon which appears when more than one contaminant are being removed at the same time. The presented model is studied and verified against experimental data.
• Author(s) : Marc Calvo-Schwarzwalder, Abel Valverde, Maria Aguareles, Timothy Myers

[13047] Energy landscape analysis for two-phase multi-component NVT flash systems by using ETD type high-index saddle dynamics

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The HiSD (High order saddle point dynamic) combined with the Rosenbrock ETD algorithm is used to calculate the solution landscape of the two-phase two-component NVT flash model based on the Peng-Robinson equation of state. The numerical results reveal the thermodynamic properties of the given system which can significantly improve the adaptability and reliability for complex engineering problems with drastic temper.
• Author(s) : Yuze Zhang,Xuguang Yang,Lei Zhang,Yiteng Li,Tao Zhang,Shuyu Sun

[13049] Dispersion of a periodically injected solute through a long circular tube

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Dispersion of a chemically reactive solute between fluid and tissue of a long circular tube is studied. Advection-diffusion equation is used for fluid flow with reversible-irreversible chemical reactions on the tube wall. Finite difference method with a time step constraint to satisfy maximum principle is used, while injection of a solute is done periodically. This kind of study can be applicable in periodic inhalation of drug.
• Author(s) : Jyoti, Soobin Kwak, Seokjun Ham and Junseok Kim

[13067] A Simple Generalized Schröter family of discrete distributions

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In actuarial mathematics, the Schröter family of discrete distributions has been analyzed using the convolution approach to determine the discrete distributions that satisfy it. However, this approach is very complex and does not work for all distributions. Thus, to address this problem, this study presents an alternative approach, a less complex derivation of the model, feasible conditions for the parameters, and estimation. Furthermore, the study presents the recursive and recursion algorithms for the derived model.
• Author(s) : Friday I. Agu

[13068] Phase Transitions in Active Polar Liquid Crystals

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We present a hydrodynamic model for an active polar liquid crystals with a variable polarization magnitude. We show that the active constituents can change the polarization dynamics and give rise to 1st and 2nd order isotropic-polar (I-P) transitions. We establish a phase diagram on the activity constituent plane.
• Author(s) : Zhenlu Cui

[13075] Global and Local Dimension Reduction to Aid Underdrawing Visualization using Hyperspectral Imaging Data from a 15th-Century Painting

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Underdrawings, preliminary sketches by an artist that are later covered with paint, are of great importance for art investigation. Utilizing shortwave infrared reflectance hyperspectral imaging to unveil these underdrawings provides spatially heterogeneous and high dimensional data, posing a challenge for underdrawing visualization. To aid visualization and complement standard PCA and the Minimum Noise Fraction transform methods, we apply new spatially-localized versions of each to Botticelli and Lippi’s “The Adoration of the Kings” (NG592, about 1472).
• Author(s) : Wallace Peaslee, Shira Faigenbaum-Golovin, Ingrid Daubechieis, Barak Sober

[13076] A three-dimensional collective cell migration model by the phase-field method

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Collective cell migration has been observed in various cellular tissues such as Drosophila eggs and cellular slime molds of D. Discoideum. The aim of this study is to develop a three-dimensional phase-field model that can represent spontaneous cell polarity formation and collective cell migrations on planar and spherical basement membranes. This presentation will show numerical simulation results about collective rotational migrations, mound formation on planer basement membrane, and elongation of spherical cellular tissues.
• Author(s) : Futa Maeda, Takamichi Sushida

[13080] Deep learning approach for segmentation of cervical arteries in CTA images

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The accurate segmentation of cervical arteries from computer tomography (CT) images is a difficult and time-consuming challenge in radiology. Its automation, however, will allow for a quantitative analysis of arterial geometrical structures in large cohort patient studies. We propose a novel framework combining segmentation methods with centerline tracking approaches. We reach state of the art accuracy while reducing inference time by a factor of approximately three compared to nnUNet.
• Author(s) : Markus Tiefenthaler, Lukas Neumann, Elke Ruth Gizewski, Stephanie Mangesius

[13081] Troubleshooting numerical simulations of PDEs using sparse regression

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We introduce a general, robust approach to data-driven modeling of spatially extended systems called sparse physics-informed discovery of empirical relations (SPIDER). The models take the interpretable form of systems of tensor-valued partial differential equations, identified using weak-form sparse regression. We demonstrate the utility of SPIDER for validating and troubleshooting results of numerical simulations. In particular, we leverage irreducible tensor decompositions to identify specific inaccurately solved equations or misspecified boundary conditions.
• Author(s) : Daniel Gurevich, Matthew Golden, Roman Grigoriev

[13086] Maximum entropy method for efficient spectrum analysis

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In R&D of optical materials such as OLED, it is required to obtain spectrum efficiently. Maximum Entropy Method (MEM) was applied to the spectrum analysis of time-series data, calculated by real-time time-dependent density functional theory, which is expensive. Compared to Fourie Transformation, MEM provided spectra of fairly high resolution with a relatively small number of time-series data. MEM is attractive in industry, because we can save significant computational costs.
• Author(s) : Yasunari Zempo, Nobuhiko Akino, Satoru S. Kano

[13087] Electronic structure calculation in meshless particle method

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The smoothed particle hydrodynamics (SPH) method is applied to the electric structure calculation in real space, described by the time-dependent Schrödinger equation. In the process of time evolution, we employed the Bohmian expression. Since the wave function is described by particles, the propagation can be directly represented by particle movement. SPH has enough accuracy and stability to use practical applications. We will report our development of SPH for the electronic structure calculation.
• Author(s) : Yasunari Zempo, Satoru S. Kano

[13089] Inferring the Utility from Optimal Behaviour in an Epidemic using Neural Networks

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We model rational individuals socially distancing in an epidemic as a differential game, where individuals seek to simultaneously maximise their own utility function by modifying their behaviour. One can solve a constrained optimal control problem to derive optimal system dynamics that result in these maximal utilities. We use Machine Learning to solve the inverse problem, to infer some unknown utility function that is being optimised by given system dynamics, assuming no knowledge of it’s form.
• Author(s) : Mark P Lynch, Matthew S Turner, John J Molina, Simon K Schnyder, Ryoichi Yamamoto

[13090] On the mathematical modelling of scale-up and intra-particle diffusion for column adsorption

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Experimental data demonstrates that the breakthrough curves of column adsorption exhibit qualitative differences related to the size of the adsorbent particles. Standard models do not account for particle size, consequently, we develop a new formulation that incorporates the diffusion of the adsorbate into the adsorbent particles. Approximate analytical solutions are developed coupled to mass flow and intra-particle diffusion equations and verified through comparison with numerical solutions of the of the full mathematical model.
• Author(s) : Abel Valverde, Alba Cabrera-Codony, Marc Calvo-Schwarzwalder, Timothy G. Myers

[13102] Optimal control of stochastic differential delay equations and their applications

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : This research is devoted to the study of optimal control of stochastic differential delay equations with jumps and Markov switchings, and their applications in economics. By using the Dynkin formula and solution of the Dirichlet-Poisson problem, the Hamilton-Jacobi-Bellman (HJB) equation and the inverse HJB equation are derived. Applications are given to two stochastic models in economics, namely, – a Ramsey diffusion model with jumps and a Ramsey diffusion model with jumps and Markov switchings.
• Author(s) : A. Ivanov, M. Svishchuk, A. Swishchuk, S. Trofimchuk

[13103] Haplotype associations with quantitative traits in the presence of confounding and population stratification.

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In genetic mapping of complex traits, when genotypes are not randomly distributed in different environments, covariance appears between genotype and environment. It is subject to the usual concerns about confounding and population stratification.
  We extend the algorithm (Shibata et al. 2010) to incorporate family-based samples.
  We describe a simple and effective method to estimate and test these haplotypic covariance, and incorporate them in to the haplotypic tests for association.
• Author(s) : Kyoko Shibata, Takashi Adachi

[13128] Test of randomness with distributions of words

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : NIST SP800-22rev1a provides nonoverlapping and overlapping template matching tests (chi-square tests for the number of
  the occurrences of words) for statistical test of random numbers.
  In this poster presentation, we demonstrate new randomness test for pseudo random numbers (LCG, Mersenne twister, etc) with KS-test and exact formulae for distributions of nonoverlapping words and runs that are recently developed by the author.
  
• Author(s) : Hayato Takahashi

[13146] Symmetries in transmission electron microscopy imaging of crystals with strain

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : TEM images of strained crystals often exhibit symmetries, the source of which is not always clear. To understand these symmetries we distinguish between symmetries that occur from the imaging process itself and symmetries of the inclusion that might affect the image. We prove mathematically that the intensities are invariant under specific transformations. A combination of these invariances with specific properties of the strain profile can then explain symmetries observed in TEM images.
• Author(s) : Anieza Maltsi

[13153] New determinant formula for $4\times 4$ matrix via Lasso

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : In this study, we propose a new determinant formula for the $4\times 4$ matrix. Observation of Derksen’s formula, sparse dictionary learning (SDL), and sparse identification of nonlinear dynamics (SINDy) inspired the determinant of the $4\times 4$ matrix to be approached from the perspective of sparse optimization. The new formula significantly improves the upperbound of the determinant tensor rank of the $4\times 4$ matrix compared to previous studies.
• Author(s) : Taehyeong Kim, Jeong-Hoon Ju, Yeongrak Kim

[13154] A New Formula of the Determinant Tensor with Symmetries

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : In this paper, we present a new formula of the determinant tensor $det_n$ for $n \times n$ matrices. Recently, Kim, Ju, and Kim found a new formula of $4 \times 4$ determinant tensor
  $det_4$ which is available when the base field is not of characteristic $2$. Considering some symmetries in that formula, we found a new formula so that $Crank(det_n) \leq rank(det_n) \leq \frac{n!}{2^{\lfloor (n-2)/2 \rfloor}}$ when the base field is not of characteristic $2$.
• Author(s) : Jeong-Hoon Ju, Taehyeong Kim, Yeongrak Kim

[13162] Generation and motion of interfaces in a mass-conserving reaction-diffusion system

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Reaction-diffusion models with nonlocal constraints naturally arise as limiting cases of coupled bulk-surface models of intracellular signalling. A minimal, mass-conserving model of cell-polarization on a curved membrane is analyzed in the limit of slow surface diffusion. Using the tools of formal asymptotics and calculus of variations, we study the characteristic wave-pinning behavior of this system on three dynamical timescales.
• Author(s) : Pearson W. Miller, Daniel Fortunato, Matteo Novaga, Stanislav Y Shvartsman, Cyrill B Muratov

[13165] Bernstein Polynomials for Finite Element Analysis

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The Bernstein basis polynomials satisfy all the necessary requirements for use as shape functions in the Finite Element Method (FEM). There are benefits to using the Bernstein polynomials over the standard Lagrange polynomials for shape functions in finite element analysis. This poster will show our investigation into p-enrichment as well as high-order elements (which produce dramatically lower condition numbers) by replacing the Lagrange polynomial basis shape functions with Bernstein basis polynomials in FEM.
• Author(s) : Oscar Alvarez, Dr. Keith Ballard*, Dr. Endel Iarve*

[13171] On sequential optimality theorems for linear fractional optimization problems involving integral functions defined on L^2n[0,1]

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We consider a linear fractional optimization problem involving integral functions defied on L^2n[0,1], and obtain sequential optimality conditions for the linear fractional optimization problem which hold without any constraint qualification, and are expressed by sequenses.By using the optimality theorems, we formulate the nonfractional dual problem for the linear fractional optimization problem, and prove the weak duality theorem and the strong duality theorem.
• Author(s) : GWI SOO KIM, MOON HEE KIM, GUE MYUNG LEE

[13181] Solving fractional disease model using new iterative method

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Fractional order derivatives and hence differential equations (FDEs) involving fractional derivatives are getting increasing attention due to their ability to model complex phenomena such as disease models which carries memory effect, hereditary properties . Systems with memory effects are the systems in which present state of the system at time t=t_0 depends on previous history times t
• Author(s) : Yogita Mahatekar, Amey S. Deshpande

[13187] Numerical simulation for generalized time-fractional Burgers’ equation with three distinct linearization schemes

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : In the present study, we examined the effectiveness of three linearization approaches for solving the time-fractional generalized Burgers’ equation using a Modified Atangana-Baleanu Caputo derivative. A stability analysis of the linearized time-fractional Burgers’ differential equation was also presented. All linearization strategies used to solve the proposed non-linear problem are unconditionally stable. Furthermore, numerical results demonstrate the comparison of linearization strategies and manifest the effectiveness of the proposed numerical scheme in three distinct ways.
• Author(s) : Reetika Chawla, Komal Deswal, Devendra Kumar, Dumitru Baleanu

[13193] Models correction based on sparse identification and data assimilation

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : When a localized damage causes a degradation of the mechanical properties of the structure, models may need suitable corrections. For instance, to reflect this local loss of performances, the stiffness matrix associated with the structure should be locally corrected. We propose an original methodology for locally correcting the models from collected data, that is at its turn completed beyond the sensors location. The technique that we propose exploits sparsity in the context of parameters identification.
• Author(s) : Daniele Di Lorenzo, Victor Champaney, Angelo Pasquale, Francisco Chinesta

[13194] Least Squares Estimation for Non-Ergodic Weighted Fractional Ornstein-Uhlenbeck Process of General Parameters

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We will investigate the least squares estimation (LSE) method for the non-ergodic weighted fractional Ornstein-Uhlenbeck (WF-OU) process with general parameters. The study focuses on deriving the asymptotic properties of LSE estimators, analyzing convergence rates, and establishing the limiting distribution. Numerical simulations demonstrate the effectiveness and robustness of the LSE method for estimating the parameters of the non-ergodic WF-OU process, contributing to a better understanding of LSE applied to complex stochastic processes.
• Author(s) : Abdulaziz al Senafi, Mishari Al-Foraih, Khalifa Es-sbaaie

[13197] Quantitative relations among causality measures with applications to pulse-output nonlinear network reconstruction

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We investigate causal connectivity in nonlinear networks with pulse signals as measured output, using four commonly used causality measures. We establish a direct link between the inferred causal and structural connectivity, demonstrating that the causal connectivity inferred by any of the four measures coincides with the underlying structural connectivity. And our pairwise reconstruction framework of structural connectivity can be achieved without conditioning on global network information, providing an effective approach for pulse-output network reconstruction.
• Author(s) : Kai Chen, Zhong-qi Tian, Songting Li, David McLaughlin, Douglas Zhou

[13199] Availability Evaluation of Warm Standby System with Fault Detection Delay and General Repair Time

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : This study analyzed the availability of a warm standby system that works with fault detection delay and general repair times. The time-to-detection delay is also considered as exponentially distributed. The detection state is used to detect the faults in the failed unit. The steady state availability of the system is obtained by using supplementary variable technique. Three types of repair time distributions are compared to find the best one.
• Author(s) : Sanjay Chaudhary, Kanta

[13201] Strong Contrasting Diffusivity with L1 source term

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The poster explains the homogenization of a second-order elliptic partial differential equation (PDE) with L1 data in a circular oscillating domain, where the oscillating part is made up of two materials with contrasting diffusivity. Due to the L1 source term, the solutions are interpreted as renormalized solutions. The polar form of variational form is used to address circular oscillations, and the primary analytical tool is the polar unfolding operator.
• Author(s) : A K Nandakumaran, Abu Sufian, Renjith Thazhathethil

[13215] Parameterization of Force Field in Molecular Simulation by Machine Learning

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Force field refers to the functional form and parameter sets used to calculate the potential energy of a system of atoms in molecular dynamic simulation. We designed the graph neural networks with new message passing algorithm and applied it to the parameterization of force field. We also proposed some new functional forms to evaluate the potential energy for different parts, which greatly improve the accuracy.
• Author(s) : Gong CHEN, Yvon MADAY

[13216] Statistical Analysis of Sled-pull Training Effects on Athletes’ Force Velocity Profiles

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The force velocity profile (FvP) is a determinant of success in soccer. This study is the first to detail FvP in both male and female collegiate Division I soccer players. We present a statistical analysis of how a 12-week sled pull training intervention affects total sprint times, maximal horizontal speed, maximal horizontal force, and maximal power output. The study helps design more effective training interventions that specifically target force and power development.
• Author(s) : John A. Brasher, Robert J. Rovetti, Junyuan Lin, Robert V. Musci, Jenevieve L. Roper

[13229] Mathematical modelling and optimal control of Zika virus with multiple interventions

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : This study presents the mosquito-borne Zika virus disease, in which impact of information-induced behavioural change, use of insecticide-treated nets, condoms, treatment, and indoor residual spraying on the vector population are analyzed. Local stability of disease-free equilibrium is established. Optimal control problem is formulated and necessary conditions are derived to minimize the infected individuals and the cost. Different combinations of control strategies are implemented and compared numerically. The combined effect of all control interventions are observed
• Author(s) : TAPAN SARKAR, SADURI DAS, PRASHANT K SRIVASTAVA, PANKAJ BISWAS

[13245] Equivariant Global Hopf Bifurcation in Abstract Nonlinear Parabolic Equations

• • Date & Time : 5P (Aug.25, 12:20-13:20)
  • Abstract : The paper proposes a method for studying symmetric global Hopf bifurcation problems in a parabolic system. The objective is to detect unbounded branches of non-constant periodic solutions that arise from an equilibrium point and describe their symmetric properties in detail. The method is based on the twisted equivariant degree theory, which counts orbits of solutions to symmetric equations, similar to the usual Brouwer degree, but on the report of their symmetric properties.
  • Author(s) : Zalman Balanov, Wieslaw Krawcewicz, Arnaja Mitra, Dmitrii Rachinskii

[13248] Theoretical and numerical analysis of the Landau-Lifshitz-Baryakhtar equation in micromagnetism

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The Landau-Lifshitz-Baryakhtar (LLBar) equation describes the evolution of spin fields in continuum ferromagnets at moderate temperature below the Curie temperature, taking into account contributions from nonlocal damping. It is a generalisation of the Landau-Lifshitz-Gilbert (LLG) and the Landau-Lifshitz-Bloch (LLB) equations in the theory of micromagnetism. Global well-posedness of the strong solution of LLBar equation is proven. Conforming finite element methods to solve this equation are rigorously analysed and implemented.
• Author(s) : Agus Soenjaya, Thanh Tran

[13259] A posteriori error estimates and their use for a least-cost strategy to achieve target accuracy

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Our work aims at providing an optimal cost strategy to achieve the targeted accuracy when approximating the solution of a nonlinear PDE. The numerical error comes from two sources: the number of iterations and the finite dimensional approximate space. We first apply a probabilistic method to explore an optimal path. Based on the analysis of this optimal path, we propose a near-optimal strategy to achieve a given accuracy based on a posteriori estimates.
• Author(s) : Muhammad Hassan, Yvon MADAY, Yipeng WANG

[13261] Application of meta-sampling method based on Maxima Weighted Isolation Kernel to the genetic data for personalized cancer care

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : An approximate Bayesian computation method based on Maxima Weighted Isolation Kernel mapping with additional parameter generation and selection processes, i.e. meta-sampling, is used to select multidimensional parameters of a personalised cancer cell evolution and cancer treatment simulator. The proposed machine learning method makes it possible to predict the effectiveness of cancer treatment for a patient.
• Author(s) : Iurii Nagornov

[13272] Hybrid Twins based on Optimal Transport

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Although experimental data is the most accurate, it is expensive to gather. Numerical simulations are cheaper but present an inherent error. Within the “Hybrid Twin” rationale, a data-driven model learns this numerical-experimental ignorance gap, being able to correct any further simulation. However, classical machine learning methods lead to non-physical results when dealing with localized solutions, as in fluid dynamics. Therefore, we propose an original ignorance model based on Optimal Transport providing a novel interpolation strategy.
• Author(s) : Sergio Torregrosa, Victor Champaney, Amine Ammar, Vincent Herbert, Francisco Chinesta

[13273] Numerical calculation of the portal pressure gradient of the human liver

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Portal hypertension refers to the abnormal increase of the portal venous pressure, which is a common chronic liver disease with clinical consequences of cirrhosis, such as hepatic encephalopathy, variceal hemorrhage and ascites. In this poster, we proposed a highly parallel method for the transient incompressible Navier-Stokes equations for the simulation of the blood flows in the full three-dimensional patient-specific hepatic artery, portal vein and hepatic vein. As applications, we also calculate the portal pressure gradient.
• Author(s) : Zeng Lin

[13284] Benchmark problems for numerical methods of the Allen-Cahn and Cahn-Hilliard equations

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We propose benchmark problems to assess the accuracy and convergence of numerical methods for phase-field models. The benchmark problem is subjected to linear stability analysis and then selected as a growth mode solution closely related to the dynamics of the governing equations. Two famous phase-field models such as the Allen-Cahn and Cahn-Hilliard equations are used to demonstrate the appropriateness of the proposed benchmark problems.
• Author(s) : Youngjin Hwang, Junseok Kim

[13285] Regularization properties of dropout gradient descent

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Generalization is a crucial aspect of training algorithms in machine learning. Dropout training has been shown to improve generalization of different models e.g. neural networks. In this work, we give a theoretical explanation of this phenomenon. We introduce a time-continuous analog of dropout gradient descent called Ornstein-Uhlenbeck dropout and study its behavior in the small noise limit. We obtain an effective limit model, in which the regularization term induced by dropout is explicit.
• Author(s) : Anna Shalova, Mark Peletier, André Schlichting

[13290] Dynamic Contracting in Asset Management under Investor-Partner-Manager Relationship

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We study incentive contracts in asset management business under dynamic actions and relationships between an investor, a partner of an investment company, and a fund manager of the company. The investor cannot perfectly observe the partner and manager’s actions, and similarly, the partner cannot perfectly observe the manager’s actions. We show how the actions of the participants and the costs of their actions interact and then extend to the case with multiple managers.
• Author(s) : Jussi Keppo, Nizar Touzi, Ruiting Zuo

[13298] Variational formulations of continuously deep neural network and existence results

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The poster discusses a mathematical analysis of a deep neural network model called ODE-Net, where learning is formulated as an optimal control problem. The existence of an optimizer for this problem needs to be assumed to justify the learning, but few studies have analyzed this in detail. We proves the existence of an optimizer when a neural network, which describes a vector field of ODE-Net, is linear with respect to learnable parameters.
• Author(s) : Noboru Isobe

[13301] Maximum principle preserving stability analysis of the fully explicit method for the Allen-Cahn equation

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : In this presentation, we analyze the stability of a fully explicit finite difference method to solve the Allen-Cahn equation. We investigate the second-order finite difference for the space and the explicit Euler scheme for the temporal derivative. The explicit method is time-step constrained, however, it is fast and accurate. We analyze and propose the time-step constraint formula preserving the discrete maximum principle and energy stability for the numerical solutions of the Allen-Cahn equation.
• Author(s) : Seokjun Ham, Junseok Kim

[13302] Numerical solution of one-dimensional Fisher–Kolmogorov–Petrovsky–Piskunov equation for unconditional stability and positivity-preserving.

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We present a new unconditionally stable and positivity preserving numerical method for one-dimensional Fisher-Kolmogorov–Petrovsky–Piskunov (Fisher-KPP) equation. The proposed method separates the Fisher-KPP equation into a diffusion equation and nonlinear equation with Operator Splitting Method. Implicit Euler method and Interpolation method is used to solve the diffusion equation and nonlinear equation, respectively. Convergence, stability, p value effects and positivity preserving is demonstrated by various numerical tests. Lastly, we introduce some possible future works.
• Author(s) : Seungyoon Kang, Junseok Kim

[13304] Rogue wave solutions to the Sasa-Satsuma equation

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In this poster, we construct rogue wave solutions to the Sasa-Satsuma equation by using the Kadomtsev-Petviashvili hierarchy reduction method and the Hirota’s bilinear method. These solutions are presented in three different forms, all expressed explicitly by rational functions with the numerator and denominator being the determinants of even order. The dynamics of the rogue wave solutions of first to third order are calculated and demonstrated.
• Author(s) : Bao-Feng Feng, Changyan Shi, Chengfa Wu, Guangxiong Zhang

[13305] Patterns of rogue waves of Long-wave and Short-wave equations

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Using the Kadomtsev-Petviashvili (KP) hierarchy reduction method, we obtained general rogue wave solutions for the Long wave Short wave system. Our approach eliminates complexity and derives rogue wave solutions through algebraic expression, resulting in diagrammed patterns up to fifth order. And N-1 polygonal patterns of Nth-order rogue waves are found to be connected to the Yablonskii-Vorob’ev polynomial hierarchy.
• Author(s) : Peng Huang, Yuke Wang,Dan Zhou

[13310] Harder-Narasimhan Filtrations of Persistence Modules

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The Harder-Narasimhan type is a discrete invariant defined on quiver representations and parameterised by central charges. We investigate its discriminative power for different settings arising in Topological Data Analysis: zigzag, rectangle-decomposable multiparameter, and nestfree ladder persistence modules. In each case, we find central charges whose Harder-Narasimhan types are complete. We also introduce a discrete invariant which is strictly stronger than the rank invariant for representations of arbitrary quivers, including multiparameter persistence modules.
• Author(s) : Marc Fersztand, Emile Jacquard, Vidit Nanda, Ulrike Tillmann

[13314] Tracking data analysis for ball possession time in football matches

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The probability distributions of player’s and team’s possession times in football matches are calculated using a large dataset.
  We find that player’s possession time follows a gamma distribution and the player number in team’s single possession follows the mixture of two geometric distributions.
  We propose a formula to express the distribution of team’s possession time using these two distributions, and confirm its validity by the data analysis.
• Author(s) : Ken Yamamoto, Seiya Uezu, Keiichiro Kagawa, Yoshihiro Yamazaki, Takuma Narizuka

[13315] Bifurcation of minimal attractor of diffeomorphism with additive and spherical bounded noise

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We introduce a novel boundary mapping to approximate and analyse bifurcation of minimal invariant set of a particular set-valued mapping which represents the support of stationary measure of a diffeomorphism with additive and spherical bounded noise. The boundary map can be used to approximate the boundary of the minimal invariant set, detect birth of singularities on the boundary and the analysis of its topological bifurcation – a discontinuous change as parameters are varied continuously.
• Author(s) : Wei Hao Tey, Jeroen S.W. Lamb, Martin Rasmussen

[13332] An epidemic dynamics model with a limited capacity of isolation for a reinfectious disease

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We consider an epidemic dynamics model with a system of ordinary differential equations for a reinfectious disease, in which a limited capacity of isolation is incorporated. By mathematical analyses on the model, we discuss the influence of the limited isolation capacity on the epidemic consequence, and try to give theoretical implications on the importance of the isolation capacity with respect to the control of a disease spread in a community.
• Author(s) : Zhiqiong FU, Hiromi SENO

[13334] Neural option pricing for the rough Bergomi model

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : This research investigates pricing financial options based on the rough Bergomi model by neural SDEs. We propose an efficient approximation of sample paths using the sum of exponentials and implement the Wasserstein distance as a loss function for network training. The option pricing is entirely based on the traditional martingale theory. Our experimental results indicate that the error of the option price can be bounded by the very Wasserstein distance attained during training.
• Author(s) : Teng Changqing

[13347] Multi-stage Stochastic control

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In stochastic networked models, different nodes may encounter different situations, so distinguishing between the distinct local controls of each node and central controls for all nodes is essential. Here we address this issue by constructing a two-stage control problem. The value functions and local controls in the first stage are optimised by fictitious play algorithm, and the second-stage value function and the central control are approximated by subgradient method.
• Author(s) : Jianxiong Sun, Harry Zheng

[13354] Social response could cause recurring epidemic outbreaks: A population dynamics model

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We consider an epidemic dynamics model incorporated the effect of social response, that is, a social collective behavior, to a disease spread in a community. Our analysis on the model with a system of ordinary differential equations implies that recurring epidemic outbreaks could be caused by such an effect of social response with the sensitivity and insensitivity to the disease spread in the community.
• Author(s) : Ying XIE, Hiromi SENO

[13355] Valuing of Timer Path-dependent Options

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Timer options are financial instruments,first proposed by SGCIB in 2007,which allow investors to exercise the options randomly under the level of volatility.We study the problem of valuing the timer path-dependent options(TPOs) where the volatility is governed by a fast-mean reverting process.Specifically,we derive analytical formulas for TPOs.Moreover,we verify the pricing accuracy of the analytic formulas of path-dependent options by comparing our solutions with the ones from the MC simulations.Finally,we experiment on the TPOs to demonstrate thepricingsensitivitieswithrespecttothemodelparameters.
• Author(s) : Mijin Ha, Donghyun Kim, Ji-Hun Yoon

[13359] Natural model reduction for kinetic equations

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : A novel framework of model reduction for kinetic equations is proposed, which employs Riemannian geometry and results in first-order symmetrizable hyperbolic equations that preserve properties including hyperbolicity, conservation laws, entropy dissipation, and linear stability. The relationship between the H-theorem for kinetic equations and stability conditions for reduced systems is discussed rigorously, determining the choice of Riemannian metrics involved in the model reduction. The framework is widely applicable to the model reduction of many kinetic models.
• Author(s) : Zeyu Jin, Ruo Li

[13361] Population dynamics model on the persistence of native species in fragmented habitat under an alien species invasion

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We consider a mathematical model of competition population dynamics incorporating the influence of habitat fragmentation with a system of ordinary differential equations. Taking account of the resource dynamics affected by the fragmentation, we will discuss the persistence of a native species under the invasion of an alien competitive species in the habitat. Some habitat fragmentation would work as a factor to sustain the persistence of native species threatened by alien species.
• Author(s) : Victor SCHNEIDER, Hiromi SENO

[13366] Mathematical Modeling of 3D Tumor Spheroid Growth Inhibition for Anticancer Drugs

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : A mathematical model describing the interaction between cancer and anticancer drugs may play an appropriate role in pharmacological research to reduce the cost of in-vitro and in-vivo experiments. In this work, we introduce an agent-based stochastic model to assess the effect of anticancer drugs on in-vitro 3D tumor spheroid growth inhibition and conduct local sensitivity analysis to evaluate the impact of parameters on the model. Numerical simulations show that the experimental data is well simulated.
• Author(s) : Yunil Roh, Jong Hyuk Byun, Il Hyo Jung

[13369] Using the Neural ODE Toolkit For a Geometric Resolution of the Numerical Sign Problem

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The numerical sign problem appears when sampling from high-dimensional integrals which oscillate rapidly in sign, causing catastrophic cancellation. Such integrals are common in quantum chemistry and condensed matter computations. One recently proposed resolution uses holomorphic flows to deform the integrands to reduce the sign oscillations. We investigate the implementation of this method, using tools from neural ODE to solve the flow and backpropagate through its integral curves.
• Author(s) : Hussain Kadhem

[13372] Data Assimilation For Quantum Nitrogen Vacancy(NV) Diamond Spectroscopy

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Nitrogen-vacancy (NV) defects in diamond have potential uses in quantum information and sensing. A statistical model was constructed for NV spectroscopy to solve inverse problems and develop a primary sensor based on its optical properties. The model considers the effects of local strain and environmental variables such as temperature, pressure, and electromagnetic fields. The Schrödinger Equation is solved to compute the theoretical spectroscopy curve, and the model’s robustness is assessed using sensitivity analysis.
• Author(s) : Dr. Tyrus Berry, Dr. Zeeshan Ahmed

[13376] Integrable Discretization of Lax dynamics of Cholesky type

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We present a integrable discretization of Lax dynamics of
  Cholesky type. The proposed dynamical system has solution,
  which is same form of solution of the continuous system.
  We focus on the associated eigenproblem, two types of time evolution of eigenvector which are normal type and splitted matrix type, five types of Lax representations, two types of parameter dynamics, two types of similarity transformation, decomposition property, solution to initial matrix which is written by eigen decomposition.
• Author(s) : Koichi Kondo and Masato Shinjo

[13378] Inferring the maturation trajectory of human iPS cell-derived megakaryocytes with topological data analysis

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Megakaryocytes are blood cells that produce platelets as they mature, and understanding their maturation process is of medical importance. We analysed single-cell gene expression data of megakaryocytes at various maturation levels using Mapper (a method of topological data analysis). Our analysis suggests that megakaryocytes may have diverse maturation pathways.
• Author(s) : Takumu Maehashi, Momoko Hayamizu

[13381] The refined error bounds for LCP of H+-matrix

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Based on the absolute value equation for minimizing two vectors, we present error
  bounds for the linear complementarity problems with an H+-matrix. Some of the
  computable bounds are given by providing the particular diagonal parameter matrix
  D. The proposed bounds improve some existing ones when D is chosen properly.
• Author(s) : Xianping Wu

[13382] Cofibrant indecomposable chain complexes parametrized by 1-dimensional posets

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Persistence theory for data analysis avoides cherry-picking parameters, studying instead how the topological features of data evolve as parameters change. To do so, it uses functors indexed by posets and valued in chain complexes. Here, we study functor categories whose indexing posets have dimension 1 , i.e. between any two comparable elements there is only one monotone path. We provide a projective model structure and characterize the indecomposable cofibrant objects.
• Author(s) : Wojciech Chacholski, Barbara Giunti, Claudia Landi, Francesca Tombari

[13383] Interaction Measures, Partition Lattices and Kernel Tests for Higher Order Interactions

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : There is increasing evidence that pairwise relationships are insufficient to model many real world systems. However, current methods for identifying higher-order interactions in data are often heuristic or fail to capture all possible factorisations of the joint distribution. Here, we introduce a family of higher order interactions that generalises to order $d$ based on lattice theory and use kernel methods to create nonparametric statistical tests of the corresponding measures for random variables and random processes.
• Author(s) : Zhaolu Liu, Robert Peach, Mauricio Barahona

[13385] Dynamical modelling of Alzheimer’s disease considered with astrocytes

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We establish a coupling model of neuronal populations and astrocyte to explore the possible mechanism of electroencephalography slowing in Alzheimer’s disease through mathmatical modelling. Dynamic nature is mainly due to changes in the amplitude and frequency of the oscillatory behavior, and can be attributed to the change of the oscillation mode caused by the limit cycle bifurcation and birhythmicity. Dysfunctional astrocyte disrupts the physiological state, causing three typical EEG slowing phenomena reported clinically.
• Author(s) : Honghui Zhang, Zhuan Shen, Lin Du

[13387] An Age-Structured COVID-19 Vaccine Roll-out Strategy in the Philippines

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Vaccination against a disease can significantly reduce mortality if properly distributed. In this work, an age-structured model of COVID-19 transmission is developed by considering three age groups: young, adult, and elderly. Respective parameter rates are determined using the COVID-19 numbers in the National Capital Region of the Philippines. Optimal control theory is employed to identify the best vaccine allocation for each age group. Results may aid decision-making towards mitigating disease transmission under limited resources.
• Author(s) : Arceo, Carlene P; Cawiding, Olivia R; de los Reyes, Aurelio V A; Escosio, Rey Audie S; Hernandez, Bryan S; Mendoza, Renier G; Mendoza, Victoria May P; Mohammad, Rhudaina Z; Salonga, Pamela Kim N; Suarez, Fatima Lois E; Sy, Polly W; Vergara, Thomas Herald M

[13388] Diffraction by a finite defect on a square lattice: an iterative Wiener-Hopf method approach

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Diffraction of a time-harmonic plane wave on various finite defects in a square lattice will be studied. This problem is reduced to a matrix Wiener-Hopf equation with exponential factors. There is no known exact solution method for these type of equations. This work aims to adapt the recently developed iterative Wiener-Hopf method to this situation. Previously it has only been applied to Wiener-Hopf equations that arise in a continuous medium setting.
• Author(s) : Elena Medvedeva, Anastasia Kisil, Raphael Assier

[13390] Dimension estimates in nonconformal graph directed iterated function systems via asymptotic perturbation

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We consider infinite graph-directed iterated function systems (GIFSs) whose contraction mappings are nonconformal. As our main result, we formulate asymptotic perturbations from conformal GIFSs to nonconformal GIFSs, and give the asymptotic behaviour of the Hausdorff dimension of the limit set of the perturbed system. We also investigate perturbed self-affine sets as special cases.
• Author(s) : Haruyoshi Tanaka

[13391] Extreme Event Projection for a Changing Chaotic Attractor

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We represent the Lorenz 1963 system as a continuous-time Markov chain. We investigate data-driven methods of defining state space partitions and deriving system statistics: generators, transfer operators, and stable-state solutions. We introduce methods to quantify uncertainty for these products. Using the Rayleigh number as a proxy for mean temperature, we investigate the evolution of the generators in a changing environment and its implication for the projection of extreme event occurrence under climate change.
• Author(s) : M. Geogdzhayeva, Andre Souza, Raffaele Ferrari

[13393] Analyzing the structure of cyclical competition using deep learning method

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We propose a new method using a deep neural network (DNN) to predict extinction in a cyclic competition model, which is motivated by the relationship between pattern formation and mobility. Numerical experiments showed that DNN performed well except for intermediate mobility, where the extinction probability increases sharply. Therefore, we suggest a hybrid method combining Monte Carlo and DNN, whose computational cost reduces remarkably compared with the conventional method.
• Author(s) : Junhyeok Choi, Bongsoo Jang

[13395] On Properties of Koopman Eigenfunctions for a Planar Singularly-Perturbed Dynamical System

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We study the Koopman operator to a planar singularly-perturbed dynamical system with an asymptotically stable limit cycle. Technically, the two principal eigenfunctions are computed for the operator associated with the fundamental frequency and the characteristic exponent (different from unity) of the limit cycle. We show that the eigenfunction associated with the fundamental frequency becomes steep on multiple manifolds of the system along the nullcline corresponding to the fast variable.
• Author(s) : Natsuki Katayama, Yoshihiko Susuki

[13396] Stuart vortices on a hyperbolic sphere

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The vortical Stuart solution of the inviscid incompressible 2D fluid flow is studied on the surface of a hyperbolic sphere with a constant negative curvature. Here, the bilinear and polynomial functions and their composites are analyzed with some illustration. The limiting solution is interpreted in terms of the point vortex flow. The similarities and differences with the planar and the spherical cases are explained.
• Author(s) : Jongbin Yoon, Habin Yim, Sun-Chul Kim

[13398] Group Inverses everywhere!

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : It is crucial to know properties and expressions for the group inverse of the Laplacian matrix in order to understand the properties fulfilled by the solution of a wide variety of problems in the discrete setting. For instance, group inverses appear in relation with discrete vector calculus, random walks, machine learning, pagerank problems, and so on. We will review some of the main properties and goodness of group inverses.
• Author(s) : Ángeles Carmona, Andrés M. Encinas, María José Jiménez

[13416] Universal Triboelectric Nanogenerator (TENG) Simulation Method and Design Automation System

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Current TENG design is still based on analytical/numerical derivation or strenuous experimental trial. A low-cost, universal and reliable design automation is not existed. An automated design method is proposed. Various randomly initiated TENG designs are calculated using finite element analysis (FEA) and equivalent circuit simulation (ECS) for subsequent two objective optimization based on NSGA-II. By introducing simplifications in FEA and ECS, the simulation can be accelerated by at least 1000 times without losing accuracy.
• Author(s) : Jinkai Chen,Hao Zhou, Jikui Luo

[13424] Hybrid Iterative Solver for Inverse Problems

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Inverse problems arise in a variety of applications: machine learning, image processing, finance, mathematical biology, and more. Solution schemes are formulated by applying algorithms that incorporate regularization techniques and/or statistical approaches. In most cases these solution schemes involve the need to solve large-scale ill-conditioned linear systems that are corrupted by noise and other errors. In this talk we consider new hybrid Krylov subspace methods to solve these linear systems, including how to choose regularization parameters.
  
• Author(s) : Ariana Brown, James Nagy

[13425] Minimizing movement for mean curvature flow with prescribed contact angle in curved domain

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We investigate a time discrete scheme for the mean curvature flow with prescribed contact angle. This is a modified version of the scheme proposed by Chambolle including a capillary functional instead of the total variation. We show that the scheme is well-defined and has consistency with the energy minimizing scheme of Almgren-Taylor-Wang type. Moreover, our solution turns out to be able to represent a strong solution to the problem in terms of Split Bregman method.
• Author(s) : Tokuhiro Eto, Yoshikazu Giga

[13429] Emergence of synchronization in Kuramoto model with frustration under general network topology

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : In this paper, we will study the emergent behavior of Kuramoto model with frustration on a general digraph containing a spanning tree. We provide a sufficient condition for the emergence of asymptotical synchronization if the initial data are confined in a half circle.
• Author(s) : Tingting Zhu

[13445] An overset grid scheme for studying particles confined to fluid interfaces

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The dynamics of particles confined to fluid interfaces are connected to the motion and geometry of the interface itself. A numerical framework is presented to study this connection for low Reynolds number flows. The approach uses an overset grid scheme to resolve the flow while interface motion is captured via the level set method. The efficacy of the method is presented for the two-dimensional case of a particle translating between parallel plates.
• Author(s) : Colton Bryant, David Chopp, Michael Miksis

[13446] The effect of “fear” on two species competition

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The fear of predation can impact competitive systems. The study analyzes ODE and PDE Lotka-Volterra competition models, where one competitor is fearful of the other, and finds that fear can have interesting dynamical effects, including novel bi-stability dynamics. The study also investigates the effects of spatially heterogeneous fear functions, showing that weak competition can turn into competitive exclusion under certain conditions. The findings have applications in ecological and sociopolitical settings.
• Author(s) : Vaibhava Srivastava, Eric M. Takyi, Rana D. Parshad

[13448] Construction of Diffeomorphism with Lagrange-Multipliers of Jacobian Determinant and Curl

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : This poster summarizes a Lagrange-Multiplier formulation of a recent work, Variational Principle (VP) that generates diffeomorphisms (non-folding grids) with prescribed Jacobian determinant and curl, for improving an optimal control approach to the task of image registration that produces VP-featured deformation fields as solutions. The improvement bypasses the explicit computation of control functions occurring in optimization while maintaining its solutions VP-featured. Such effectiveness is demonstrated with numerical examples.
• Author(s) : Zicong Zhou, Guojun Liao

[13449] Time-domain Maxwell’s equations in biperiodic structures

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We focus on the mathematical analysis of the diffraction of an electromagnetic plane wave by a biperiodic structure. The method of a compressed coordinate transformation is proposed to reduce equivalently the diffraction problem in a bounded domain over a finite time interval. The reduced problem is shown to have a unique weak solution by using the constructive Galerkin method. The stability and a priori estimates with explicit time dependence are established.
• Author(s) : Jue Wang

[13451] Physics Informed Neural Networks for Vibration Equations of Large Membranes Arising in the Music Industry

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The new paradigm in ML focused on combining scientific principles with ML algorithms; known as scientific machine learning (SciML). In this article, we have implemented a SciML algorithm, viz. Physics Informed Neural Networks, to solve vibration equations of large membrane (VeLM). VeLM has great importance due to its broad applications in areas such as music, biomechanics and acoustics. To prove the reliability of model, the numerical measures are presented using error analysis and solution graphs.
• Author(s) : Arup Kumar Sahoo, Sandeep Kumar and S. Chakraverty

[13461] Improving constraint stability of covariant BSSN formalism of the Einstein equations against homogeneous and isotropic spacetime background

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : In solving the Einstein equations numerically, the BSSN formalism is more widely used due to its numerical stability.
  However, constraint violation cannot be avoided even with the BSSN formalism, depending on the background spacetime.
  We consider the stability of the Einstein equations against homogeneous and isotropic spacetime background in the covariant form of the BSSN formalism. We propose that adjusting the time evolution equation with constraints contributes to improving the constraint stability.
  
• Author(s) : Hidetomo Hoshino, Takuya Tsuchiya, Gen Yoneda

[13462] Improving EIT discrete techniques

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : For health issues, Electrical Impedance Tomography (EIT) represents a non-invasive and radiation-free imaging technique for recovering the conductivity distribution inside the body under observation from skin surface measurements. Besides, EIT is known to be a groundbreaking area of research because its low cost and portable instrumentation. It is well known that this problem is is severely ill-posed. Therefore, new algorithms to overcome this structural difficulty are necessary.
• Author(s) : María José Jiménez Jiménez, Ángeles Carmona, Andrés M. Encinas

[13470] Analysis of solidification phenomena in Bulkley-Herschel extrusion flows

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Simplified flow models for Bulkley-Herschel fluids in ceramic paste extrusion are investigated. Exact solutions are derived analytically and discussed. In particular, it is demonstrated that for a certain range of model parameters there exist solidification zones. The effects of the model parameters on solidification-zone formation are rigorously analyzed. Furthermore, the analogy to the widely investigated formation of dead zones in chemical reactor engineering is shown.
• Author(s) : Piotr Skrzypacz, Bek Kabduali, Rustem Takhanov, Vsevolod Andreev, Boris Golman

[13471] Novel Semi-Analytical Methods for Nonisothermal Diffusion-Reaction Processes in Catalyst Pellets with Arbitrary Reaction Kinetics

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Solutions to problems of non-isothermal diffusion-reaction in catalyst pellets with external mass and heat transfer limitations, as well as arbitrary reaction kinetics, are developed in the form of Taylor series. Novel semi-analytical and closed-form explicit approximations are derived for the concentration of reactants and the temperature in catalyst pellets of various geometries. The proposed methods are verified numerically for both isothermal and non-isothermal problems with power-law kinetics.
• Author(s) : Boris Golman, Vsevolod V. Andreev, Piotr Skrzypacz

[13472] Multigrid POD Galerkin Method for multiscale inhomogeneous PDEs

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Model Order Reduction through POD-Galerkin method gives low-order expressions of the solution to PDEs. Solving multiscale PDEs with POD, however, faces the problem of deriving spatial basis mixed with different frequency. A revision mPOD combines information from both energy and frequency perspective. It splits solution into contributions from different frequency. Galerkin method is then applied to the downsampled low frequency contributions for higher efficiency. The improvement is a large constant times faster than conventional methods.
• Author(s) : Tianhao HU ; Zecheng GAN

[13473] Achieving High Accuracy with PINNs via Energy Natural Gradient Descent

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We propose energy natural gradient descent, a Hessian-based natural gradient method as an optimization algorithm for physics-informed neural networks. We show that the update direction in function space resulting from the energy natural gradient corresponds to the Newton direction modulo an orthogonal projection onto the model’s tangent space. We demonstrate experimentally that energy natural gradient descent yields highly accurate solutions with errors several orders of magnitude smaller than standard optimizers.
• Author(s) : Marius Zeinhofer, Johannes Müller

[13474] Topologies on unparameterised path space and signature asymptotics

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The signature of a path is a non-commutative exponential, initially investigated by K.T. Chen in the 1950s, which has found recent success in data science. One of its key properties is the ability to uniformly approximate continuous functions on compact subsets of unparameterised paths. We present recent results on the choice of topology on the space of unparameterised paths, as well as a generalisation of a theorem of Hambly-Lyons concerning the asymptotics of signature terms.
• Author(s) : William F. Turner, Thomas Cass, Remy Messadene

[13488] Free Vibration of Multiferroic Laminates with Interfacial Imperfections and Nonlocal Effect

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : A semi-analytical procedure is elaborated for the free vibration of multiferroic laminated plates with interface imperfections and nonlocal effect. The wave motion characteristics such as dispersion curves and mode shapes are investigated. We extend the pseudo Stroh formalism and the propagator matrix method to derive the solutions for the proposed configurations. The derived solutions are then applied to BaTiO₃-CoFe₂O₄ sandwich plates. Numerical calculations show that these effects have a pronounced influence on wave motion characteristics.
• Author(s) : Hsin-Yi Kuo, Po-Chun Huang

[13489] Heuristic algorithm for finding Hamiltonian cycles in random geometric graphs

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Random geometric graphs (RGGs) have been recently applied as a model of various real-world spatial networks. While the theoretical aspects of monotone properties of RGGs have been studied, the relevant computations are less considered. In this work, we develop a heuristic algorithm for finding Hamiltonian cycles in RGGs in different domains. Experimental results are presented to demonstrate the effectiveness of the algorithm and shed light on the close relationship between connectivity and Hamiltonicity in RGGs.
• Author(s) : Mandy Man

[13490] Adaptive discrimination between harmful and harmless antigens in the immune system by predictive coding

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The immune system discriminates between harmful and harmless antigens based on past experiences; however, the underlying mechanism is largely unknown. Here, we modeled the population dynamics of T cells by adopting the concept of predictive coding into immunological memory formation. Through numerical simulations, we found that the immune system identifies antigen risks depending on the concentration and input rapidness of the antigen. Further, our model reproduced history-dependent discrimination, as in allergy onset and subsequent therapy.
• Author(s) : Kana Yoshido, Honda Naoki

[13491] Constrained Optimization Using Philippine Eagle Optimization Algorithm

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The Philippine Eagle Optimization Algorithm (PEOA) is a recently proposed population-based single-objective bound-constrained optimization algorithm based on the hunting behaviors of the Philippine Eagle. To make PEOA a more general optimizer, this study proposes a version of PEOA that can solve constrained optimization problems. We conduct experiments to test its performance in solving benchmark functions against other constrained algorithms. We also use the method in real-world constrained applications, including COVID-19 vaccine allocation in the Philippines.
• Author(s) : Erika Antonette Enriquez, Renier Mendoza, Arrianne Crystal Velasco

[13492] A transform approach to multi-phase problems

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We describe a novel approach to multi-phase problems, based on transform methods, allowing for the construction of explicit solutions and expedient calculation of integral quantities. This method is exemplified by application to the leading order analysis of a three-dimensional two-phase problem inspired by liquid-infused surfaces, providing formulae for flows, pressure gradients and slip lengths. Further utility is demonstrated by quantifying ‘shear-driven failure’ in such a surface, patterned with microscopic grooves. [Joint work with Darren Crowdy]
• Author(s) : Henry Rodriguez Broadbent, Darren Crowdy

[13493] Stochastic properties of pairs in hand of playing cards having arbitrary number of suits

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We investigate stochastic properties of the number of pairs in a random subset of a playing-card deck whose number of suits is not limited to four. We derive the closed forms of the average and variance of pairs using the Gauss hypergeometric function. Moreover, the covariance and correlation coefficient between pair numbers in two subsets are calculated.
  
• Author(s) : Satoshi Umeki, Ken Yamamoto

[13495] Characterizing Eigenspaces of Fisher Information Matrix in Simple ReLU Networks

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We investigate the Fisher Information Matrix (FIM) of one hidden layer neural networks with ReLU activation, focusing on the weight vector from the hidden layer to the scalar output. We characterize the first three clusters of the eigenvalues and eigenvectors of the FIM, using the row vectors of the weight matrix of the previous layer. Our results shed light on the geometry of the FIM eigenspaces, providing insights into learning behavior in neural networks.
• Author(s) : Yoshinari Takeishi,Masazumi Iida,Jun’ichi Takeuchi

[13497] Shape optimization of an isolated body in incompressible viscous flow for minimization of drag force based on Deep Q-Network and the FEM

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We daily use the vehicles running in the fluid such as planes and railways. The
  performances of these are heavily depended on the aerodynamics represented on drag
  and lift. Moreover, the forces are determined by the machinery form. Therefore, to find
  an optimal shape is important problem in industrial field. In this research, we
  approached this problem with a machine learning called DQN as a fundamental
  experiment, and clarified it is valid for limited condition.
• Author(s) : Soma Hirooka, Yudai Sugiyama, Takahiko Kurahashi

[13498] Application of density-based topology optimization for maximally stiff structure problem using two-phase materials

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In this study, numerical experiments for maximally stiff structure problems using two-phase materials are conducted by density-based topology optimization, and the strain energy is compared for cases with and without separation for two-phase materials. The MBB model is employed in the computational model, and some numerical results will be shown by changing numerical conditions.
• Author(s) : Mizuki Ikarashi, Masayuki Kishida, Takahiko Kurahashi

[13499] Texture shape optimization for minimization of friction coefficient based on the modified accelerated gradient method

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In this study, we present a shape update formula of optimization to reduce friction on the
  bearing surface. In this optimization, a shape update formula has been proposed by
  applying the Taylor expansion to the gradient descent method. Nesterov’s accelerated
  gradient method is known to have better convergence than the gradient descent method.
  The Taylor expansion is applied to the Nesterov’s accelerated gradient method to improve
  the convergence.
• Author(s) : Hideto Oda, Takahiko Kurahashi

[13500] Level-set-based topology optimization for bi-linear type elasto-plastic problems

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The purpose of this study is to develop a topology optimization method that combines the manufacturability and reliability to solve the problem of weight reduction in the automotive industry. In this study, a method that combines the level-set and phase-field methods was applied to a bi-linear elasto-plastic model, which can control the complexity of the geometry. The results show that support components are generated to reduce strain energy at stress concentration regions.
• Author(s) : Kouske Shimizu, Masayuki Kishida, Takahiko Kurahashi

[13501] A robust lower order mixed finite element method for a strain gradient elasticity model

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : A robust lower order nonconforming mixed element method in two and three dimensions is developed for a strain gradient elasticity (SGE) model. At the beginning, the uniform regularity of the SGE model is derived under two reasonable assumptions. Then, the robust discrete inf-sup condition is established, and the sharp and uniform error estimates are achieved with respect to both the small size parameter and the Lam'{e} coefficient. Numerical results verify the theoretical findings.
• Author(s) : Mingqing Chen, Jianguo Huang, Xuehai Huang

[13502] Second-order flows for computing the ground states of rotating Bose-Einstein condensates

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We introduce two types of dissipative second-order hyperbolic partial differential equations as computational models for numerically approaching the ground-states of rotating Bose-Einstein Condensates. The proposed artificial dynamics are motivated from recent progress of accelerated gradient flows in convex optimization, though constrained non-convex variational problems are facing here. Numerical schemes for the proposed dynamics provide several novel, efficient and robust algorithms to the field of BEC ground-state simulation, which will be sketched in the poster.
• Author(s) : Haifan Chen, Guozhi Dong, Wei Liu, Ziqing Xie

[13503] Neural network stochastic differential equation models with applications to financial data forecasting

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In this work, we employ a collection of stochastic differential equations with drift and diffusion coefficients approximated by neural networks to predict the trend of chaotic time series which has big jump properties. Then, we theoretically prove that the numerical solution through our algorithm converges in probability to the solution of the corresponding stochastic differential equation. Moreover, we illustrate our method by applying it to real financial time series data and find the accuracy increases.
• Author(s) : Luxuan Yang,Ting Gao,Yubin Lu,Jinqiao Duan,Tao Liu

[13506] Analysis of desiccation crack pattern formations based on physics-informed neural networks

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : This study proposes a physics-informed neural network model for shrinkage-induced cracks due to drying observed on the surface of thinly spread dense colloidal suspensions (desiccation cracks) to analyze the elementary physical process underlying the pattern formation.
• Author(s) : Shin-ichi Ito

[13508] Hybridizable Discontinuous Galerkin Methods for Magnetic Advection-Diffusion Problems

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We introduce and analyze hybridizable discontinuous Galerkin (HDG) methods for solving magnetic advection-diffusion problems. Specifically, we demonstrate the stability of the method and provide corresponding priori error estimates by introducing a weight function for the more general Friedrichs system. Additionally, we propose local postprocessing techniques to enhance the accuracy of the computed vector field. To validate the effectiveness of the HDG method, we provide extensive numerical examples.
• Author(s) : Jindong Wang, Shuonan Wu

[13509] A comprehensive model for viscoplastic fluids in superhydrophobic channels

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Viscoplastic fluid flows in channels equipped with a groovy superhydrophobic surface are studied, through comprehensive mathematical modeling and complementary numerical simulations. The viscoplastic rheology is captured using the Bingham model, while the wall superhydrophobicity is established through arrays of stick and slip conditions. An arbitrary direction for the pressure gradient with respect to the groove orientation is considered, allowing to address the flow physics for longitudinal, transverse and oblique groove configurations.
• Author(s) : Rahmani, Hossein; Taghavi, Seyed Mohammad

[13510] Variational treatment of boundary value problems in Cosserat elasticity

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We consider a variational formulation of several interesting problems arising in Cosserat elasticity. Using the boundary integral equation method, we derive analytical solutions in a weak (Sobolev) space setting and show their application to the solution of practical engineering problems. As an example, we consider stress concentrations around rigid inclusions and cracks embedded in an infinite domain in the plane case. We also discuss the approaches for the treatment of the problems with irregular boundaries.
• Author(s) : Stanislav Potapenko

[13512] PIEZO1 regulates leader cell formation and cellular coordination during collective cell migration

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The mechanically-activated ion channel PIEZO1 was recently identified to play an inhibitory role during wound healing. Through an integrative experimental and mathematical modeling approach, we elucidate PIEZO1’s contributions to keratinocyte collective migration, an essential component of the healing process. Here, through a 2D-multiscale model of wound closure which links observations at both the single and multicell scales, and subsequent experimental validation, we identify cell directionality as being impacted by PIEZO1 activity during wound closure.
• Author(s) : Jesse Holt, Jinghao Chen, Elizabeth Evans, John Lowengrub, Medha Pathak

[13519] Variance-Reduced Stochastic Subspace Descent

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Many optimization problems, such as hyper-parameter learning and PDE constrained optimization, have costly objective functions for which the gradient is unavailable. We propose a novel descent method, Variance Reduced Stochastic Subspace Descent (VRSSD), which uses randomized directional derivative approximates combined with control variance to produce low-variance gradient estimators. We demonstrate convergence of VRSSD and illustrate similar empirical performance to a gradient descent benchmark using as few as ten percent of the number of function evaluations.
• Author(s) : Killian Wood, Emiliano Dall’Anese, Stephen Becker

[13524] Numerical Evaluation of Mixed Precision Iterative Refinement using Low Precision Krylov Methods.

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Mixed precision (MP) numerical methods exploiting low precision computing have attracted attention under the recent computational hardware trend. In this study, we investigate MP iterative refinement based solvers using typical Krylov subspace methods, namely GMRES and BiCGSTAB, in FP32. Experimental results show the MP solvers can provide solutions with the same accuracy as solvers using only FP64. Comparison results of execution time of the MP solvers and conventional solvers using FP64 are also given.
• Author(s) : Yingqi Zhao, Takeshi Fukaya, Takeshi Iwashita

[13528] Phoneme-guided speech separation by using non-negative matrix factorization

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : This study proposes a method for speech separation using non-negative matrix factorization (NMF) to categorize two speakers in audio processing. The method exploits phoneme similarity and does not require pre-training. It considers neighboring phonemes to determine the speaker after separation. Results show that the separation accuracy can be boosted by adjusting the sample rate, which is also analyzed in the study
• Author(s) : June-Ho, Lee

[13529] Using a machine learning approach to forecast severe cases of COVID-19

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Early identification of severe cases of COVID-19 patients is an important issue. We predict moderate and severe cases using machine learning algorithms based on COVID-19 data visiting the emergency room of a hospital. We show the pre-processing processes of the data, the predictive performance results and the feature importance corresponding to the machine learning models. We compare the machine learning training results with the results of determining the severity of patients in the hospital.
• Author(s) : Jung Eun Kim, Tobin Kim, Sunmi Lee, Hee-Sung Kim

[13530] Long short term memory based stock price predictions

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Stock price prediction is a challenging task due to the complex and dynamic nature of financial markets. In this paper, we propose a Long Short-Term Memory (LSTM) based model for stock price prediction. We collect historical stock price data, preprocess it, and train an LSTM model using a suitable architecture. We evaluate the model’s performance using common metrics and fine-tune it if necessary. Our results demonstrate the effectiveness of the LSTM model in predicting stock
• Author(s) : Dilpreet Kaur and Kavita Goyal

[13531] Artificial neural network (ANN) based derivative pricing

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Derivative pricing is a crucial task in financial markets, requiring accurate and efficient models. In this poster, we propose an ANN based approach for derivative pricing. We leverage the power of ANNs, specifically feedforward neural networks, to capture complex nonlinearities in the derivative pricing process. We discuss the data preparation, model architecture, and training process for our ANN model. We also present results demonstrating the potential of our approach in accurately pricing derivatives.
• Author(s) : Dilpreet Kaur and Rohit Kumar Singla

[13536] Value-Gradient Based Formulation of Optimal Control Problem and Machine Learning Algorithm

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : In this work, instead of focusing on the value function, we propose a new formulation for the value-gradient as a decoupled system of partial differential equations in the context of a continuous-time deterministic discounted optimal control problem. An efficient iterative scheme for this system of equations is also developed by combining the characteristic line method with machine learning techniques. Experimental results demonstrate that this new scheme significantly increases the accuracy and improves the robustness.
• Author(s) : Alain BENSOUSSAN, Jiayue HAN, Sheung Chi Phillip YAM, Xiang ZHOU

[13538] The LP-Newton Method with Separation Hyperplanes for Linear Programming

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : In this talk we discuss the LP-Newton method of linear programming, an algorithm that differs from the well-known methods such as Simplex method and Interior Mthoed. This method can be attributed to finding the intersection of a polyhedron and a line. The purpose of this presentation is to propose a new algorithm for finding the intersection of polyhedra and lines and to verify its efficiency by numerical experiments under appropriate assumptions.
• Author(s) : Matsuno Yuki , Jianming Shi

[13549] Mathematical representation of bias and nudge centered on intangible goods by quantum information theory

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The purpose of this study is to explore whether the relationship between biases and nudges can be mathematically expressed in terms of quantum information theory, particularly with individually customized nudges. I propose a model of bias and nudge that takes into account the environment for intangible goods. This increases the feasibility of the economic model based on quantum information and the mathematical design of customized nudges.
• Author(s) : Misao Fukuda

[13551] Predicting Dengue Hemorrhagic Fever Incidents in DKI Jakarta by Considering Climate Factors using Machine Learning Models

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The number of Dengue Hemorrhagic fever cases has increased more than eightfold over the past two decades. Prediction of future DHF incidents involves machine learning, whether by deep-learning (convolutional neural network, extreme gradient boosting, Long Short-Term Memory, Transformer) or non-deep learning (LSTM-ATT, Poisson Regression, support vector regression). Data is daily DHF incident data and climate data in DKI Jakarta from 2008-2023. The resulting prediction can prevent the number of cases from increasing.
• Author(s) : Bevina D. Handari, Fairuzia Zahira, Anuriyah Pebriana, Hilmi T. Shalahudin, Naufal Alfarisi, Gianina Ardaneswari, Dipo Aldila

[13552] Identifying Sensitive Areas for Targeted Observations to Improve Air Quality Prediction over South Korea

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In air quality prediction, initial conditions, based on atmospheric and aerosol/chemistry observations, are essentially required for a coupled atmosphere-chemistry model. Therefore, to improve air quality prediction, it is essential to find the upstream areas of the region of interest. The conditional nonlinear optimal perturbation for initial conditions (CNOP-I) is a suitable method to identify the sensitive regions. We introduced the key variables and sensitivity area for aerosol by using the principal component analysis and CNOP-I.
• Author(s) : Seungyeon Lee, XiaoHao Qin, Jiwon Yoon, Seon Ki Park

[13553] A faster prediction-correction framework for solving convex optimization problems

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : He and Yuan’s prediction-correction framework [SIAM J. Numer. Anal. 50: 700-709, 2012] is able to provide convergent algorithms for solving convex optimization problems at a rate of $O(1/t)$ in both ergodic and pointwise senses. This paper presents a faster prediction-correction framework at a rate of $O(1/t)$ in the non-ergodic sense and $O(1/t^2)$ in the pointwise sense, {without any additional assumptions}. Interestingly, it provides a faster algorithm for solving {multi-block} separable convex optimization problems.
• Author(s) : Tao Zhang

[13554] A family of Barzilai-Borwein steplengths from the viewpoint of scaled total least squares

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The Barzilai-Borwein (BB) steplengths play great roles in practical gradient methods for solving unconstrained optimization problems. Motivated by the observation that the two well-known BB steplengths correspond to the ordinary and the data least squares, respectively, we present a family of BB steplengths from the viewpoint of scaled total least squares. Numerical experiments demonstrate that high performance can be received by a carefully-selected BB steplength in the new family.
• Author(s) : Shiru Li, Tao Zhang, Yong Xia

[13556] Network formation by replicating coupled oscillators

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The assembly of small cellular networks – graphs of cells coupled via cytoplasmic bridges – remains poorly understood despite its biological importance. To address this, we develop a minimal model of network formation based on coupled oscillators. Each oscillator replicates after completing a cycle, enabling network growth. By modifying three key parameters, the model synthesizes data from a broad range of naturally occurring cellular networks, thus establishing a foundation for theoretical extensions and experimental validation.
• Author(s) : Matthew Smart, Stanislav Shvartsman, Hayden Nunley

[13557] Cooperative Security Against Interdependent Risks

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Firms in inter-organizational networks such as supply chains or strategic alliances are exposed to interdependent risks. These risks can be decomposed into intrinsic risks a firm faces from its own operations and extrinsic risks transferred from its networked partners. We examine whether and when firms can cooperatively secure themselves against such risks via cost-sharing arrangements that are stable (i.e., belong to the core), fair (in a formalizable sense) and implementable via bilateral cost-sharing arrangements.
• Author(s) : Sanjith Gopalakrishnan, Sriram Sankaranarayanan

[13558] Data-based Optimal Tuning of I-PD Controllers for Time-Delayed Unstable Processes

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : This study presents a data-based method under the virtual reference feedback tuning (VRFT) framework for optimal tuning of I-PD controllers to control time-delayed unstable processes. An optimization problem is formulated to determine an appropriate reference model for the controlled process with explicit considerations of output performance, control effort, and system robustness. The controller tuning can be dealt with by using a multi-objective optimization algorithm to generate a set of trade-off optimal solutions (Pareto front).
• Author(s) : Jyh-Cheng Jeng, Wen-Jeng Chen

[13560] Modeling Surface Tension in a Multi-Material Numerical Framework

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Pacific Island Structured-amr with ALE (PISALE) is a multiphysics simulation framework built on a multi-material arbitrary Lagrangian-Eulerian numerical solver, used for a wide range of applications. We describe the implementation of a surface tension package in PISALE using a slightly modified but conventional volume-of-fluid approach and a height function to approximate the boundary. The accuracy, robustness, and efficiency of the package are discussed, as well as applications to modeling droplet breakup in high repetition-rate experiments.
• Author(s) : Jack McKee

[13566] Bayesian physics-informed neural networks for seismic tomography based on function-space particle-based variational inference

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Existing approaches that combine Bayesian inference with physics-informed neural networks (PINN) for PDE-based inverse problems have limitations in the problem size and uncertainty estimation accuracy. In this study, we propose a new approach to Bayesian PINN using particle-based variational inference in the function space instead of the weight space. We find that the proposed method is computationally feasible with accurate results for seismic tomography, a class of inverse problems which estimates subsurface seismic wave velocity.
• Author(s) : Ryoichiro Agata,Kazuya Shiraishi,Gou Fujie

[13568] Investigation of Barren Plateaus in Quantum Neural Network Development

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Training of Quantum Neural Networks (QNN) inherits the flaws of Machine Learning itself, which includes the Barren Plateau that impedes the convergence of Gradient-Based optimisers.
  This project, therefore, aims to investigate the effectiveness of different approaches to mitigate the emergence of barren plateaus in various QNN developmental circumstances.
  To this end, three different approaches were selected to deal with barren plateaus and were then evaluated against different VQA quantum circuit structures (depth, qubits and initialisation), initially using a random gradient landscape and then on a sample classification problem.
• Author(s) : Nguyen Ngo Cong Thanh, Jacob Cylbulski

[13570] Ensemble-based data assimilation system for satellite aerosol observation and regional aerosol prediction model

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Ensemble-based data assimilation (DA) was advantageous for assimilating aerosol observations because of the ensemble-estimated flow-dependent background error covariance. In this study, we perform DA using the Maximum Likelihood Ensemble Filter (MLEF), one of the ensemble-based DA methods, and the Weather Research and Forecasting model with Chemistry (WRF-Chem) on an Asian Dust Storm case in East Asia. Assimilation of satellite observed aerosol optical depth improves WRF-Chem prediction skill.
• Author(s) : Ebony Lee, Milija Zupanski, Seon Ki Park

[13571] Distributionally Robust Crew Pairing

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Airline operations are subject to various uncertainties, which usually reduce operational efficiency and cause extra crew costs. To mitigate the impact of uncertain factors, we propose a distance-based distributionally robust crew pairing model via Kullback-Leibler divergence, considering minimizing the expected cost under the worst-case delay distributions restricted in an ambiguity set. Meanwhile, we also develop an exact and tractable solution approach which can successfully solve industry instances in a reasonable time and provide promising results.
• Author(s) : Di Xiao

[13576] Fast L sigma method for Variable order Fractional Derivative

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : This work proposes a fast and high-order numerical scheme compromising a shifted binary block partition (SBBP) algorithm and the Lσ method for solving Caputo variable-order(VO) time-fractional diffusion equation. Improving the Lσ method for the VO fractional derivative, this novel approach has a convergence order of O(∆3−\bar{α}) with a computational complexity to O(n log n), where \bar{α}=||\alpha(t)||_{\infty}.
• Author(s) : Junseo Lee

[13577] Uncertainty quantification on process variability for magnetic tunnel junction (MTJ)/ CMOS hybrid logic circuits

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Process variability is ubiquitous in chips manufacturing and may lead to lower yield. As modern technology integrates ever more transistors, standard practice such as Monte Carlo Simulation is rendered too expensive to model the impacts of process variability. In this poster, we propose a stochastic collocation method coupled with designed quadrature rules for high-dimensional parameters. Its overall performance in convergence rate and costs is tested and verified in magnetic tunnel junction/CMOS hybrid logic circuits.
• Author(s) : Yue Zhang, Peng Wang

[13578] Impact of viral load on COVID-19 transmission dynamics considering the variants

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We aim to predict epidemiological periods considering viral load of SARS-CoV-2 using a mathematical model. To do so, first, we identified the factors affecting the viral load. Then we analyzed viral load data for Delta and Omicron variants. The parameters are estimated using the statistical methods. Then we constructed a model with those parameters to predict viral dynamics according for the Delta and Omicron variants. Finally, we examined impact of viral load on transmission dynamics.
• Author(s) : Eunseo Choi, Hyojung Lee

[13579] Estimating the early detection of COVID-19 outbreak using machine learning

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We aim to analyze transmission patterns of COVID-19 outbreak and to detect the start timing of new outbreak of COVID-19. We investigate important features that affect the start time of COVID-19 outbreak. We develop a new approach that combines machine learning methods with mathematical and statistical methods. As the result, the new outbreaks are detected from the suggested method with the high accuracy more than 90% in detecting the transmission trend.
• Author(s) : Hyeonjeong Ahn, Giphil Cho, JeongRye Park, Yongin Choi, Hyojung Lee

[13580] State Estimation for a High-dimensional Nonlinear System by Particle Smoothing Method

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In this study, the performance of particle smoothing method in nonlinear system state estimation is verified. The smoothing methods is applied to a relatively high-dimensional ocean model conceptually representing the Atlantic thermohaline circulation. The numerical results show that the particle smoothing method exhibits superior capability for state estimation of nonlinear systems, and show how the smoothing method can help improve the estimation of the filtering.
• Author(s) : Meiyan Jiang, Sangil Kim

[13582] Multi-attention based recurrent neural network for hand-foot-mouth disease prediction in Korea

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Hand-foot-mouth disease (HFMD) is a viral disease that mainly affects children. It is prevalent every year in Korea and is greatly affected by meteorological factors. In this study, a new HFMD prediction model is proposed using a multi-attention based recurrent neural network (MA-RNN). We predict suspected cases of HFMD in children under 5-years-old by week and find that humidity and amount of sunshine group played an important role among meteorological factors.
• Author(s) : Sieun Lee, Soyeon Kim, Sangil Kim

[13585] A Phylogenetic Analysis of Migratory and Resident Birds

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Migratory birds change their habitat with the seasons, while resident birds remain in one place throughout the year. The factors underlying these differences are not fully understood. In this study, we design a distance function to measure the dissimilarity between different birds and construct phylogenetic networks from the distances using the Neighbor-Net algorithm. Our analysis reveals important factors that distinguish migratory and resident birds.
• Author(s) : Yukino Kawai, Tatsuya Hisada, Momoko Hayamizu

[13587] Phylogenetic Analysis of Flapping and Soaring in Birds: Uncovering Evolutionary Differences

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Birds are divided into two types of flight style: flapping and soaring. The specific factors that distinguish them remain unknown. We calculated a distance matrix between birds from various public data on avian ecology and morphology, and visualized the distance information using phylogenetic reconstruction methods. The results suggest several important factors that differentiate bird flight styles.
• Author(s) : Tatsuya Hisada, Yukino Kawai, Momoko Hayamizu

[13588] Constructing a Phylogenetic X-cactus from a Distance Matrix

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Phylogenetic trees are the standard model used to represent the evolutionary process of organisms. Algorithms for constructing trees from distance matrices are widely used. However, phylogenetic networks provide a more powerful model of evolution for describing reticulate events such as hybridization. In this poster, focusing on a special type of phylogenetic networks called phylogenetic X-cactuses (a.k.a. level-1 networks or galled trees), we present an algorithm for computing a phylogenetic X-cactus from a distance matrix.
• Author(s) : Keita Watanabe, Momoko Hayamizu

[13590] The high-speed scaling and squaring for the matrix functions appeared in exponential integrators

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We consider the algorithm to evaluate the matrix $\varphi$ functions. The well known algorithm is scaling and squaring that combines some approximation formula and the double-angle relations. However, due to many matrix products in the relations, the optimum degree of the formula has not been considered. So we propose the new method that is equivalent to the relations and does not require many matrix products. Furthermore, we obtain the optimum degree of the approximation formula.
• Author(s) : Shinsuke NAKAMURA

[13591] Early Detection of Norovirus Outbreaks Using Machine Learning

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The norovirus is a significant contributor to acute gastroenteritis at any age. Because of this, it’s critical to detect norovirus outbreaks. We proposed a new risk index that takes meteorological characteristics, oyster production, and previous detection rate, etc. We used machine learning techniques to estimate the norovirus epidemic based on the risk index. Additionally, we determined the feature importance for the risk potential to lead to norovirus outbreak.
• Author(s) : Geunsoo Jang

[13593] Teaching computational inverse problems: what can ChatGPT do?

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Recently, there has been a lot of talk about the effect of ChatGPT on assignments. At University of Helsinki, we teach inverse problems with a computational view and have exercises and an exam the students solve without supervision. Preliminary testing the assignments on ChatGPT show an acceptable answer when explaining concepts. However, when asked to provide code for the computational part, the code gives an error message after running it.
• Author(s) : Heli Virtanen

[13595] Disease severity and time to severity prediction using deep learning and survival modelling

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In clinical practice, ideally we would like to predict timing and severity grading of a disease but we often do not see the natural course of the disease as intervention is made once it is detected. We develop a methodology using deep learning and survival modelling to predict the time to incidence of different grades of disease, allowing for more personalized treatment profiles.
• Author(s) : Yichen Chen,Sören Dittmer,Michael Roberts

[13596] Constructing data-driven ODEs of a chaotic fluid flow

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We construct ODEs describing the macroscopic variables of a chaotic fluid flow from time-series data.
  With the method to make ODEs from scalar time-series data, we construct a system of ODEs that predicts the high-frequency energy time series of the fluid flow from the low-frequency energy time series.
  The constructed system is evaluated by not only short time series but also the invariant sets.
• Author(s) : Natsuki Tsutsumi, Yoshitaka Saiki, Kengo Nakai

[13597] Mathematical modeling of chromatin dynamics in Hox-mediated animal body development

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Hox genes are core factors for animal development. These genes are linearly arrayed on the genome and expressed sequentially along the body axis, known as “collinearity”. It is still unclear how continuous upstream signals are converted into a sequential expression. To address this question, we formulated a thermodynamical model for the chromatin dynamics around Hox genes. It revealed the sequential activation of Hox genes, suggesting a thermodynamical mechanism of the Hox collinearity.
• Author(s) : Yoshifumi Asakura, Yoshihiro Morishita, Takayuki Suzuki

[13599] Iterative Linear Solvers for Interior Point Methods with Applications in Radiation Therapy

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We study the use of iterative linear solvers for interior point methods in the context of quadratic programming for radiation therapy treatment planning. We implement a prototype interior point method using iterative linear solvers, and evaluate it on real-world problems. For our application area, we find that an iterative linear solver can be used, even with a simple preconditioner to give solutions in reasonable time, though the number of Krylov iterations can become large.
  
• Author(s) : Felix Liu, Albin Fredriksson, Stefano Markidis

[13600] Counting the trees inside a phylogenetic network: an analytic combinatorial approach

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The trees underlying a phylogenetic network have been extensively studied. Recently, an algorithm has been developed to count specific spanning trees in a network in linear time of the input network size. However, there are still many open problems related to this counting problem, such as counting the number of non-isomorphic and non-homeomorphic spanning trees in a network. This poster discusses these unsolved problems using an analytic approach based on the generating functions.
• Author(s) : Hiroaki Kojima, Momoko Hayamizu

[13601] Inheritance Pattern of Crosses of Yellow Mice, a Multiplayer Game

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : This study will show how evolutionary game theory may be used in the analysis of the inheritance pattern of crosses of yellow mice based on experiments of Cuènot, Castle and Little. Using Gokhale and Traulsen’s model, the payoff matrix of a four-player game representing interactions between yellow and non-yellow alleles will be simplified into a two-player game. Equilibrium points and evolutionarily stable strategies of this game will be determined through replicator dynamics.
• Author(s) : Genrev Josiah Villamin, Yvette Fajardo-Lim

[13602] Comparisons of sensitive areas identified by adjoint sensitivity, singular vector, and conditional nonlinear optimal perturbations for tropical cyclone targeted observations

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : This study investigates the differences among the sensitive areas identified by methods of adjoint sensitivity (ADS), singular vector (SV), and conditional nonlinear optimal perturbation (CNOP) for typhoon targeted observation. For typhoons with weak nonlinearity, the sensitive areas identified using the three methods were similar, whereas for typhoons with strong nonlinearity, the sensitive areas identified by the LSV and ADS methods were similar, but they were quite different from those identified by the CNOP method.
• Author(s) : Feifan Zhou, Yiwei Ye, Wansuo Duan, He Zhang

[13603] Deep Levy Processes for Financial Modeling

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Levy processes are a popular tool in modeling jump-diffusion processes common in the financial literature. This work will detail a new multiply subordinated Levy process model, which we term a deep variance gamma process, including techniques for parameter estimation, depth selection of subordination, and rigorous theoretical results. We also test our estimation technique on a set of intraday one-minute cryptocurrency returns, illustrating that our approach outperforms other state-of-the-art subordinated Levy process models.
• Author(s) : Caitlin Berry, William Kleiber

[13604] Investigating the impact of weather variables to dengue incidence in the Philippines

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Determining how key weather factors influence dengue fever is vital in crafting strategies for intervention in the Philippines where the disease incidence is high. In this work, a general model-based causal inference method is applied to time series data to infer the causal relationships between weather variables and dengue cases. Results show that there is significant combined regulation from both rainfall and temperature when time delay is considered.
• Author(s) : Olive R. Cawiding, Se Ho Park, Aurelio A. de los Reyes V, Jae Kyoung Kim

[13605] Computational Modelling of Tritium Transport in the Anterior Segments of the Ocular Globe

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The cataract-prone lens of the eye is among the most radiosensitive tissues in the body. New occupational eye-lens dose limits underscore the need for accurate dosimetry prediction, including from tritium intakes, to demonstrate regulatory compliance. The proposed computational model considers multiple physical phenomena and physiological transients responsible for tritium’s transport from the blood to the eye lens by means of aqueous humor circulation. Resulting dosimetric implications originate from tritium delivery and its corresponding residence time.
• Author(s) : L. Ivan, M. Chu, D. Beaton, A. Hanu, J. Atanackovic, J.G. McDonald

[13607] An optimal control perspective on diffusion-based generative modeling

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We establish a connection between generative modeling based on SDEs and three classical fields of mathematics, namely stochastic optimal control, PDEs and path space measures. Those perspectives will be both of theoretical and practical value, for instance allowing to transfer methods from one to the respective other field or leading to novel algorithms for sampling from unnormalized densities. Further, the connection to HJB equations leads to novel loss functions which exhibit favorable statistical properties.
• Author(s) : Lorenz Richter, Julius Berner

[13609] Important nonlinear temperature advection responsible for the asymmetrical amplitude of El Nino and La Nina.

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The role of individual component of nonlinear temperature advection (NTA) in the amplitude asymmetry of ENSO is ambiguous. From the view of optimal growth of initial perturbation, this study explored which component of NTA plays important roles in ENSO amplitude asymmetry. The results indicate that the zonal and meridional NTAs enhance the intensity of El Nino, and the vertical NTA weakens the intensity of La Nina. The combined effects cause the amplitude asymmetry of ENSO.
• Author(s) : Xu Hui, Duan Wansuo, Mu Mu

[13611] A fast CO2 storage simulator using reduced bases

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : CO2 storage is a computationally intensive numerical simulation in practice, because engineers analyze the sensitivity of a complex PDE-based model to the numerous physical parameters.
  It is possible to reduce the computational time of such “many-query simulations” with the help of the Reduced-Basis method adequately deployed in a goal-oriented context.
  Reliable error bounds are key to the reduction process.
• Author(s) : Jana Tarhini, Quang Huy Tran, Guillaume Enchéry, Sébastien Boyaval

[13612] Ensemble Kalman filtering with an alternative representation of uncertainty

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In high-dimensional non-linear uncertainty quantification problems, much of the uncertainty is epistemic in nature, i.e. it comes from a lack of knowledge rather than from random perturbations. We introduce a version of the ensemble Kalman filter based on a dedicated representation of epistemic uncertainty, demonstrate its performance on both linear and non-linear filtering problems and show how it helps formulating standard heuristics such as localisation in a principled way.
• Author(s) : Chatchuea Kimchaiwong

[13613] New symmetric-hyperbolic PDEs for viscoelastic fluids

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We propose a new system of PDEs to model 3D viscoelastic flows of Maxwell fluids.
  Our system, quasilinear and symmetric-hyperbolic, unequivocally models smooth flows on small times, while ensuring propagation of waves at finite-speed.
  Our system rigorously unifies fluid models with elastodynamics for compressible solids,
  and it can be extended for applications in environmental hydraulics (shallow-water flows) or materials engineering (non-isothermal flows).
• Author(s) : Sébastien Boyaval

[13614] Numerical Approach to the Near Field Refractor Problem

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The near field refractor problem involves determining an interface (lens), between two media of propagation of light, that is capable of refracting a light beam of a given illumination intensity emanating from a punctual source so that the refracted rays form a prescribed intensity distribution to illuminate a given target set. A mathematical model to the problem and an iterative method of constructing the interface by using ovoids as building blocks is obtained.
• Author(s) : Cristian Gutierrez , Henok Mawi

[13615] SIR Endemic Model in Semi-Markov Media

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : This research is devoted to the development of averaging and diffusion approximation principles for the Endemic SIR model in semi-Markov media. The population under investigation is divided into subgroups with different rates of disease spreading. Randomness is introduced through the coefficients of the model: the coefficients are directed by a semi-Markov process which serves as a switching process. Contacts between subgroups are modeled through the use of transition probabilities.
• Author(s) : Mariya Svishchuk

[13616] Integral equations within the framework of supervised learning

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We explore the relevance of a novel approach to supervised learning with integral equations (joint work with Prof. Dr. Ralf Kornhuber and Dr. Patrick Gelß) for a further understanding of neural networks. In an attempt to shed some light on the issue of generalization from a new perspective, we present some interpretations of our numerical results.
• Author(s) : Aizhan Issagali

[13619] Multidimensional generalization of phyllotaxis obtained from products of linear forms

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Spiral patterns obtained by the golden angle method are known as phyllotaxis (a model of leaf arrangement) in botany. How to generalize the golden angle methods has been an open problem recently raised by several mathematicians. We report that by using the theory of products of linear forms, it can be applied to general surfaces and general dimensions. The developed theory is transversely related to Markoff theory, manifolds with diagonalizable metrics, Lamé partial differential equations.
• Author(s) : Ryoko Oishi-Tomiyasu

[13620] Newtonian vs. non-Newtonian effects on predictions of left atrial hemodynamics

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Atrial fibrillation, an arrhythmia characterized by chaotic electrical signals diminishing the heart’s contractility, leads to inefficient pumping that may result in clot formation, increasing the stroke’s risk. Assuming Newtonian behavior of blood is common in arterial blood flow simulations, primarily based on high shear rates. We compare Newtonian vs. non-Newtonian atrial blood flows, showing that the Newtonian rheology assumption may lead to an overestimation of shear rates and velocities in regions of stagnant flow.
• Author(s) : Sergio Nabil Gadur, Cécile Daversin-Catty, Kristian Valen-Sendstad

[13621] Spatiotemporal integral kernel of a hierarchical differential equation model in vision and afterimage phenomena

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : In our study, we propose a differential equation model based on the cellular interactions of retinal nerve cells arranged hierarchically. Next, we derive the explicit form of spatiotemporal integral kernel and show that the shape of integral kernel is constructed by the temporal biphasic function and the Mexican hat function, as shown in experimental observations of the retinal nerve cells. Finally, we give a theoretical prediction for visual mechanisms about appearance of afterimage phenomena.
• Author(s) : Shintaro Kondo, Masaki Mori, Takamichi Sushida

[13626] A Numerical Scheme for Fractional Partial Differential Equation Based on Green Function and CAS Wavelets

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : This study manifests to acquire an efficient novel numerical scheme named Green-CAS to obtain the accurate numerical solutions of the fractional-order partial differential equations. This approach is not only simple and easy to implement due to the Green function, but it also vanishes operational matrices for boundary conditions. To show the validity of the recommended technique, the order of convergence for two parameters has been demonstrated and the acquired outcomes are compared with renowned techniques.
• Author(s) : Muhammad Ismail, Bongsoo Jang

[13629] Assessing the potential impact of repellent use, early screening, and vector control on lymphatic filariasis transmission

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We study a model of filariasis transmission under repellent use, early screening, and vector control use. We show that these interventions have a good potential to eliminate filariasis. However, the basic reproduction number is not the only threshold. The critical quality of the treatment has an important role in the lymphatic filariasis elimination program. It is possible to have a stable endemic equilibrium even if the basic reproduction number is less than one.
• Author(s) : Dipo Aldila, Joseph P. Chavez, Sheryl N. Salim

[13634] Backward bifurcation and permanence of a disease-severity-structured epidemic model with limited treatment capacity

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : When an unknown pathogen is encountered, in addition to developing medicines and vaccine to counter its effects, medical collapse caused by limited treatment capacity must be seriously considered. To discuss the capacity effects on disease transmission, we present an epidemic model with necessary treatment only for severely infectives. We demonstrate the occurrence of backward bifurcation for relatively small treatment capacity, which suggests disease outbreak can reach a high endemic level under a usual disease control.
• Author(s) : Yasuhisa Saito, Hiromu Gion

[13636] Deflated CRS and BiCRSTAB methods with preserved duality

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : By imposing the orthogonality condition to the deflation subspace for deflated BiCG on BiCR, and by applying the improved preconditoning for product-type methods proposed by Itho et al, deflated CRS and BiCRSTAB methods are derived with duality preserved. The effect of deflation is observed in these algorithms as well as in the CGS and BiCGSTAB methods.
  In addition, the explicit deflation requires less computational cost for the initial shadow residual.
• Author(s) : Shuhei, Takaya

[13638] Principal Geodesic Analysis applied to path signature

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Principal Geodesic Analysis (PGA) is applied to path signature. First, we transform each multidimensional sequence into the path signature, and treat the signature space as a Lie group with canonical Cartan connection. Then, the group mean, and the first and principal components are derived by solving stationary value problems. The efficacy of the method is demonstrated using a climate time series.
• Author(s) : Nozomi Sugiura

[13643] Bijective Density-Equalizing Maps for Multiply Connected Open Surface

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : This paper introduces an algorithm for computing bijective density-equalizing mappings on multiply connected open surfaces in $\mathbb R^3$. Existing methods only focus on simply connected surfaces or 2D domains. The algorithm uses quasi-conformal flow and Beltrami coefficients to ensure bijectivity and preserve geometry structure. The proposed method enables effective computation of the flattening mapping from multiple connected surfaces to planes and can be extended for landmark registration. Experiments demonstrate its effectiveness.
• Author(s) : Zhiyuan LYU, Lok Ming LUI

[13645] Intervention Strategies for the Reduction of Smoking in Adolescents

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Tobacco smoking remains a major public health threat worldwide. Many adult smokers are a problem, but so are the numbers of youth smokers. Studies show that teens are relatively more likely to share smoking and drinking habits with their peers. In our research, it mathematically modeled on youth smoking based on the fact that youth smoking is contagious.
• Author(s) : Byul Nim Kim, Chunyoung Oh

[13648] Convergence Rates of Stochastic Zeroth-order Gradient Descent for Łojasiewicz Functions

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We prove convergence rates of Stochastic Zeroth-order Gradient Descent (SZGD) algorithms for Lojasiewicz functions. The SZGD algorithm iterates as $x_{t+1}=x_t-\eta_t\hat{\nabla}f(x_t)$, $t=0,1,2,3,\cdot,$ where $f$ is the objective function that satisfies the Łojasiewicz inequality with Łojasiewicz exponent $\theta$, $\eta_t$ is the step size (learning rate), and $\hat{\nabla}f(x_t)$ is the approximate gradient estimated using zeroth-order information only.
  Our results show that $f(x_t)$ can converge faster than $\{\|x_t-x_∞\|\}_{t\in\mathbb{N}}$, regardless of whether the objective $f$ is smooth or nonsmooth.
• Author(s) : Tianyu Wang, Yasong Feng

[13649] Application of Threshold Scheme to Anonymous Public-key Certificate

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : A threshold scheme is a typical scheme in secret sharing schemes.
  The most famous one is Shamir’s scheme (1979).
  Takeuti and Adachi (2023+) propose a new threshold scheme using Latin squares.
  On the other hand, Oishi et. al. (1998) proposed anonymous public-key certificates, which can guarantee anonymity of the certificate users.
  In this poster, we talk about application of the new threshold scheme using Latin squares to anonymous public-key certificates.
• Author(s) : Tomoko Adachi, Kazuomi Oishi

[13650] A Multilayer Level Set Method for Modelling Dynamic Interfaces with an Application to an Elliptic Inverse Problem

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The level set method is a numerical method for computing the motion of a single interface. This work considers the simultaneous evolution of multiple interfaces. We will introduce an extension to the original level set method, called the multilayer level set method. Then we will look into concrete examples of dynamic interface and inverse problems.
• Author(s) : Shingyu Leung, Ken K.T. Hung

[13651] Assessing Excess Mortality During the COVID-19 Pandemic in South Korea

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : The COVID-19 pandemic has had significant impact worldwide. To estimate excess mortality during the pandemic, we used monthly death and mean temperature data from South Korea. Our analysis revealed significant regional variation in excess mortality, with some areas experiencing higher rates of excess deaths not attributed to COVID-19. Targeted interventions and public health measures are needed to address indirect effects of the pandemic on mortality in these areas, protecting vulnerable populations.
• Author(s) : Sunhwa Choi, Soyoung Kim

[13653] Exponential rates of convergence for binary classification with Deep Neural Networks

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : We study the binary classification problem under Tsybakov’s strong low noise condition, which characterizes the fact that the two classes are “well separated” from one another. Without relying on the overly restrictive Neural Tangent Kernel regime, we prove that mildly overparametrized ReLU Neural Networks enjoy an exponential convergence rate of their excess risk with respect to the number of samples under this assumption, bridging the gap between theoretical guarantees and practical observations for these models.
• Author(s) : Nathanael Tepakbong

[13654] A Comparative study between CNN, CNN-LSTM for video classification

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The CNN-LSTM model is a neural network that combines a convolutional neural network (CNN) model and a long short-term memory (LSTM) model to effectively learn spatiotemporal features of image data. Experiments using UCF public data, wave height data, etc. show that the CNN-LSTM model has higher predictive performance for a given data set. Therefore, it shows advantages when making predictions on time-dependent data such as video data.
• Author(s) : Miji KIM, Sangil KIM

[13658] On degree of illusional effect of different three-dimensional objects reconstructed from the same line drawings

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : When people see objects, they recognize them as two-dimensional information on the retina. As a result, the image is a “line drawing” with missing depth information, and people recognize it as a three-dimensional object by reconstructing it from the line drawing. Given depth information, different three-dimensional objects can be reconstructed from the same line drawing. In this presentation, we calculate “degree of illusional effect” using the vertex data and report its characteristics.
• Author(s) : Takanori Asaki , Akiyasu Tomoeda

[13659] Preconditioned twisted factorization method with cyclic reductions for computing eigenvectors

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : The twisted factorizations and cyclic reductions are techniques for solving equations of linear systems. The twisted factorizations also have an application to computing eigenvectors of symmetric tridiagonal matrices. In this paper, we propose a hybrid method based on the twisted factorizations and cyclic reductions for symmetric tridiagonal eigenvectors.
• Author(s) : Masato Shinjo, Kojiro Sato, Masashi Iwasaki, Yoshimasa Nakamura

[13660] Exploring Heterogeneity in Serial Intervals of COVID-19 in South Korea

• Date & Time : 5P (Aug.22, 12:20-13:20)
• Abstract : This study aims to examine the heterogeneity in serial intervals of COVID-19, which vary by age and region in South Korea. The Korea Centers for Disease Control and Prevention (KCDC) provided the epidemiological data from January 2020 to July 2021. We extracted the empirical serial intervals between infector-infectee pairs. Our analysis includes examining age-specific and region-specific serial intervals over time, as well as investigating the effects of different interventions on serial intervals.
• Author(s) : Tobhin Kim, Hyosun Lee, Hyoeun Kim, Sunmi Lee

[13661] LONG-TIME BEHAVIOR OF COMPOSITE WAVE OF PLANAR VISCOUS SHOCKS FOR THE 3D BAROTROPIC NAVIER-STOKES EQUATIONS

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We will present the latest result on long-time behavior of solutions to the 3D compressible Navier-Stokes equations with initial $H^2$-small perturbation of Riemann data. Especially, we consider the Riemann data generating Riemann solution composed of planar shocks with small amplitudes. We prove that the solution of Navier-Stokes system converges, uniformly in space, towards a composition of two planar viscous shock waves as time goes to infinity, up to dynamical shifts.
• Author(s) : Lee Ho Bin

[13662] Tuning Convergence Rate via Non-Bayesian Social Learning: A Trade-Off between Internal Belief and External Information

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Inspired by the diversity of agents when they are exposed to new information, we design a non-Bayesian learning strategy, named as Parametric Social Learning, by introducing an agent stubbornness parameter to trade-off the significance in between its internal belief and external information. This strategy thus allows for tuning convergence rate by adjusting the introduced parameter, which is consistent highly with the sociological intuition. Theoretical analyses and numerical examples are provided to illustrate several sociological insights.
• Author(s) : Dongyan Sui, Chun Guan, Zhongxue Gan, Wei Lin, Siyang Leng

[13663] Investment and consumption under stochastic risk premia

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : We aim to solve an optimal investment-consumption problem in a market comprised of n-risky assets wherein all model parameters are driven by a separate k-dimensional diffusion process. Our approach is to solve a corresponding variational problem whose solution is equivalent to solving the HJB equation from the stochastic control approach of the original optimisation problem. We focus mainly on the complete market case while providing ideas for future work on the incomplete market case.
• Author(s) : Emmet Lawless, Paolo Guasoni

[13664] Numerical Simulation of the 2020 Taal Volcanic Ash Dispersion

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The Taal Volcano underwent a phreatic eruption in 2020. During times like this, quick and data-driven decisions are needed to ensure people’s safety and minimize damages. This work simulates the volcanic ash dispersion during the eruption using the atmospheric dispersion model and estimates the diffusion rate using two approaches. The estimated diffusion rate offers insights into the extent and potential trajectory of volcanic ash dispersion, which is essential for risk assessment and disaster management.
• Author(s) : Arrianne Crystal T. Velasco, Rhudaina Z. Mohammad, Renier G. Mendoza, Ken Matthew C. Oliva

[13665] A New Way of Showing Ambiguous Objects Using Refraction

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : “Ambiguous Objects” proposed by Prof. K. Sugihara are a type of impossible object, which is an illusionary object that appears to have a completely different shape when viewed from two different viewpoints. His works have been attracting many people due to the surprise in the change of their appearance. In this paper, we propose a new way of showing ambiguous objects by considering the refraction of light to achieve the two viewpoints.
• Author(s) : Yuzu Hanaki, Akiyasu Tomoeda

[13667] A new traffic flow model described by a delay partial difference equation with bistability

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : 　In this research, we focus on bi-stability, which is a general feature empirically observed in a traffic flow, and propose a new nonlinear difference equation with the time delay effects that describes a traffic flow.
  We have performed a bifurcation analysis to obtain the range of bi-stability and found that the controllable parameters in our model have the potential to eliminate traffic jams from the perspective of stability.
• Author(s) : Kazuya Okamoto, Tomoyuki Miyaji, Akiyasu Tomoeda

[13669] Walking Direction-Based Method for Smooth Pedestrian Movement Using 3D Point Cloud Data

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : 3D point cloud data has the advantage of being able to take measurements regardless of weather and time of day, and of protecting personal information.
  Although many studies have been reported on pedestrian detection and tracking, methods for controlling pedestrian flow based on tracking data have not yet been fully established.
  In this presentation, we propose a method to achieve smooth pedestrian movement by focusing on the variations in walking direction of pedestrians.
• Author(s) : Sora Kurihara, Akiyasu Tomoeda

[13670] User-friendly Dashboard for Estimating the Risk of Emerging Infectious Diseases using a Delayed Stochastic Model and Exploring Optimal Response Strategies

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : The COVID-19 pandemic emphasized the importance of predicting and preparing for emerging infectious diseases. This study implements a delayed stochastic model based on SEIR to simulate an emerging infectious disease outbreak before any interventions, providing a user-friendly dashboard for users to set parameter distributions and review simulation results. The median outcome of the stochastic simulation is then used to identify the optimal intervention policy through an evolutionary algorithm and cost calculations.
• Author(s) : Jongmin Lee, Eunok Jung

[13671] Existence and uniqueness of weak solution to Navier-Stokes equations in 3D up to activation

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : We prove long-time and large-data existence of weak solution for large class of initial- and boundary value problems concerning three-dimensional flows of incompressible fluids. A robust mathematical theory is developed for unsteady internal flows of fluids characterized by implicit constitutive equations in the bulk and on the boundary. Finally, we focus on the
  question of uniqueness of such weak solutions, while Navier-Stokes problem with activation is a special case in this class.
• Author(s) : Miroslav Bulicek, Josef Malek, Erika Maringova

[13674] CURVATURE ALIGNED SIMPLEX GRADIENT

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : We introduce the Curvature Aligned Simplex Gradient (CASG) for gradient estimation of noisy black-box functions. CASG is the simplex gradient that minimizes the mean-squared-error using a second-order approximation of the true function. We show that surprisingly, unlike finite difference schemes, the difference vectors are non-orthogonal. CASG significantly outperforms existing methods using the same number of function evaluations and approaches the performance of central differences using only half the number of function evaluations.
• Author(s) : Daniel Lengyel, Panos Parpas

[13678] Multiplying the absolute values from Student’s t-distribution of two degrees produces the Bradley-Terry model

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Consider the Bradley-Terry model (1952) where each player i has a latent value πᵢ > 0 to “beat” another player j with the probability πᵢ / (πᵢ + πⱼ). If one supposes “i beats j” occurs in this model iff πᵢ |vᵢ| > πⱼ |vⱼ| with independent random variates vᵢ, vⱼ ~ D where D represents a probability distribution, a solution of D is, interestingly, Student’s t-distribution with 2 degrees of freedom.
• Author(s) : Toshiyuki Shimono

[13681] From samples to persistent stratified homotopy types

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : The natural occurrence of singular spaces in application has led to recent investigations on performing topological data analysis (TDA) in a stratified framework. Here, we present how to associate to point cloud data a so called persistent stratified homotopy type, an invariant suited to keep track of and detect singularities. Furthermore, we provide theoretical guarantees for our method, in particular a sampling theorem for (certain) Whitney stratified spaces with two strata.
• Author(s) : Lukas Waas, Tim Mäder

[13683] Parameter Estimation in Neutral Delay Differential Equations Using Genetic Algorithm with Multi-Parent Crossover

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Neutral Delay Differential Equations (NDDEs) have many applications in modelling physical and biological systems. Various techniques are studied to estimate the parameters of such models, one of which is using heuristic algorithms. In this work, Genetic Algorithm with Multi-Parent Crossover (GA-MPC) was applied to solve the parameter estimation problems in NDDE models with discrete delays. Results showed that GA-MPC consistently provided good fits to the data, demonstrating its potential for parameter estimation in NDDE models.
• Author(s) : Cristeta U. Jamilla, Renier G. Mendoza, Victoria May P. Mendoza

[13687] Appropriate Combining Methods for East Asian Aerosol Optical Thickness Data

• Date & Time : 3P (Aug.23, 12:20-13:20)
• Abstract : Aerosols affect on numerous aspects, not only the problem of human health, but also radiative systems of the earth. Moreover, the many aerosol products, mainly represented by the aerosol optical depth (AOD), are abundantly provided. However, the source and shape of AOD data are also enormous, which hinder proper handling of the AOD data. Here we investigate the best way to composite multi-source data, such as ground based, satellite based and merged data set.
• Author(s) : Hae Soo Jung, Seon Ki Park

[13688] Data assimilation for estimating change points of time-varying reproduction numbers

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : In this work, we propose a sequential data assimilation scheme that updates estimates of the time-varying reproduction number (Rt) of infectious diseases as new data becomes available. By inferring change points, abrupt shifts in transmission dynamics and changes under interventions can be automatically identified. We evaluate our method using real and synthetic data under multiple scenarios. The results have implications for public health decision-making, such as identifying periods of heightened transmission and informing control strategies.
• Author(s) : Han Yong Wunrow, Sen Pei, Jeffrey Shaman, Marc W. Spiegelman

[13690] Stochastic Optimization of Operational Flexibility of Hydroelectricity for Renewable Integration

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Different from wind and solar power generation, hydroelectricity has not only zero emission and low generation cost, but also high controllability and operational flexibility, enabling hydroelectric units to start up and increase their power output quickly to prevent power supply shortage risk resulting from disruptive events. This talk will adopt data-driven optimization approaches to demonstrate the flexibility of hydroelectricity, and its benefits for large-scale renewable energy integration to modern power systems.
• Author(s) : Neng Fan

[13694] Lie Symmetry Analysis of Ramani Equations

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : Shock waves propagating in a plasma medium are modeled by the sixth-order nonlinear Ramani equation. All possible closed-form solutions to the Ramani equation are derived through the holistic approach using the Lie symmetry analysis. This involves computing Lie point symmetries and the corresponding similarity reductions. However, certain symmetry-reduced ordinary differential equations possess no point symmetries. We investigate such equations through singularity analysis. Their solutions are obtained as right Painlev{‘e} series.
• Author(s) : Rajeswari Seshadri, Sherin Agnus

[13696] Steady Streaming and Pumping Driven by Two Frequency Oscillations

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : Recent experiments showed net transport of an object undergoing two mode vibrations for certain frequency pairs. Inspired by these experiments, we revisit the classical problem of steady streaming in fluids (i.e. nonzero mean flow produced from periodic forcing) driven by multifrequency oscillations. Our numerical simulations show that net pumping occurs when these two frequency pairs produce a non-antiperiodic driving force. Furthermore, we use small amplitude analysis to understand the mechanism that leads to pumping.
• Author(s) : Hyun Lee, Robert D. Guy, William D. Ristenpart

[13707] An immersed peridynamics method for fluid-driven material damage and failure in biomaterials

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : An immersed peridynamics method simulates fluid-structure interaction along with material damage and failure by using peridynamics.
  In this work, finite element-based mesh generation techniques are adopted for non-uniform material discretization that can capture complex geometries.
  The constitutive correspondence model enables the modeling of anisotropy in biomaterials.
  Non-failure numerical tests demonstrate that the proposed method yields comparable accuracy to the conventional continuum-based methods.
  We also validate fluid-driven fracture mechanics in realistic biomaterials.
• Author(s) : Keon Ho Kim, Boyce E. Griffith

[13729] Towards R(5,5)

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : For the lower bound improvement of the diagonal Ramsey number R(5,5), we constructed a 2-color coloring of the complete graph of 42 vertices with a total of only 2 complete graphs of 5 vertices of the same color.
• Author(s) : Masanori Yamanaka

[13729] Towards R(5,5)

• Date & Time : 4P (Aug.24, 12:20-13:20)
• Abstract : For the lower bound improvement of the diagonal Ramsey number R(5,5), we constructed a 2-color coloring of the complete graph of 42 vertices with a total of only 2 complete graphs of 5 vertices of the same color.
• Author(s) : Masanori Yamanaka

[13734] Effective and Efficient Neural Operator for High-Dimensional Partial Differential Equations

• Date & Time : 5P (Aug.25, 12:20-13:20)
• Abstract : Neural operator is a promising tool for operator learning that, which maps between infinite-dimensional spaces. We develop an efficient and effective algorithm for high-dimensional partial differential equations(PDE). In this poster, we will show the algorithm and its experimental results.
• Author(s) : Shintaro Fukushima

[13733] A Lipschitz Bandits Approach for Continuous Hyperparameter Optimization

• Date & Time : P (Aug., 12:20-13:20)
• Abstract : We introduce BLiE: a Lipschitz-bandit-based algorithm for Hyperparameter Optimization (HPO) that only assumes Lipschitzness. BLiE exploits the landscape of the objective function to adaptively search over the hyperparameter space. Theoretically, we establish performance guarantee for BLiE. Empirically, we demonstrate that BLiE outperforms the state-of-the-art HPO algorithms on benchmark tasks. We also apply BLiE to search for noise schedule of diffusion models. Comparison with the default schedule shows that BLiE schedule imporves sampling efficiency.
• Author(s) : Yasong Feng, Weijian Luo, Yimin Huang, Tianyu Wang

[13924] Randomization of Spectral Risk Measures and Distributional Robustness

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : In this paper, we randomize the SRM by introducing a random parameter in risk spectrum. When the distribution of the random parameter is unknown, we propose a distributionally robust formulation of the RSRM. We discuss in detail computational schemes for solving the optimization problems based on the RSRM and the distributionally robust RSRM. Finally, we discuss how to use step-like approximation and SAA to approximate the model, and derive error bounds to justify the approximations.
• Author(s) : Li Manlan, Tong Xiaojiao, Xu Huifu

[13925] Global optimization for the portfolio selection model with high-order moments

• Date & Time : 2P (Aug.22, 12:20-13:20)
• Abstract : In this paper, we study the global portfolio optimization. It is a kind of portfolio selection model with high-order moments. We introduce a perturbation sample average approximation of the reward function, which leads to a robust approximation of the new nonconvex portfolio optimization model. The approximated problem can be solved globally with Moment-SOS relaxations. We summarize a semidefinite algorithm and give numerical examples to show the efficiency of the algorithm.
• Author(s) : Yang Liu, Yang Yi, Zhong Suhan