Registered Data

[CT124]

[01205] Finite-Time Analysis of Single-Timescale Actor Critic

  • Session Date & Time : 5D (Aug.25, 15:30-17:10)
  • Type : Contributed Talk
  • Abstract : Actor Critic (AC) is a popular class of reinforcement algorithms. Despites its remarkable success in practice, the convergence property of AC is largely unexplored. Most of the existing convergence results are derived based on either double-loop update or two-timescale step sizes rule. In this talk, we presents the latest sample complexity result of practical single-timescale AC algorithm, where the actor and critic are updated alternatively with step sizes being of the same order.
  • Classification : 68T42, 90C26
  • Author(s) :
    • Qijun Luo (The Chinese University of Hong Kong Shenzhen)
    • Xiao Li (The Chinese University of Hong Kong, Shenzhen)

[01045] Advances on intimate partner femicide applying machine learning techniques and algorithms

  • Session Date & Time : 5D (Aug.25, 15:30-17:10)
  • Type : Contributed Talk
  • Abstract : Intimate partner femicide (IPF) is female’s leading cause of violent death worldwide. The accuracy of existing risk assessment instruments for IPF developed by conventional statistics is not completely competitive. In this study, machine learning techniques and algorithms for classification were used to discriminate between lethal or non-lethal violence against women by intimate partners and to detect which variables discern them the most. The obtained evidenced-based knowledge could assist professionals in predicting and preventing lethality.
  • Classification : 68Txx
  • Author(s) :
    • Esperanza Garcia-Vergara (Universidad Loyola Andalucia)
    • Carlos Fresneda-Portillo (Universidad Loyola Andalucía (Spain))

[02091] Stock Data has Shape: Managing Stock Portfolio via Topology-informed Machine Learning

  • Session Date & Time : 5D (Aug.25, 15:30-17:10)
  • Type : Contributed Talk
  • Abstract : Given time-series data of a stocks portfolio, we generate sequences of point-cloud embeddings and use topology-based features to train a classifier on a binary classification task: determine whether or not a stock performs well on day-to-day trading. Using the trained classifier, we predict which stocks in our portfolio are projected to earn on a future trading day. We evaluate our topology-informed classifier via standard metrics and projected cumulative earnings based on a tier-structured investment scheme.
  • Classification : 68Txx, 68Uxx, Machine Learning
  • Author(s) :
    • Paul Samuel Ignacio (University of the Philippines Baguio)

[00031] Stoneley wave in the transversely isotropic thermoelastic diffusion materials

  • Session Date & Time : 5D (Aug.25, 15:30-17:10)
  • Type : Contributed Talk
  • Abstract : This paper investigate the secular equations of Stoneley wave propagating through bonded and unbonded interfaces between two dissimilar transversely isotropic thermoelastic diffusion half-spaces. These equations are solved numerically by modelling two distinct crustal rocks and the resultant phase velocities and attenuations are plotted graphically. This analysis explicates the position and permeability of fractures and also helps in the assessment of valuable materials under earth’s crust. Some particular cases are also deduced from the present formulation.
  • Classification : Continuum Mechanics
  • Author(s) :
    • Sanjay Debnath (Mizoram University)

[01115] Torsional surfce wave in a piezoelectric fiber-reinforced composite layer in context of surface/interface theory

  • Session Date & Time : 5D (Aug.25, 15:30-17:10)
  • Type : Contributed Talk
  • Abstract : The present paper investigates surface/interface theory-based torsional surface wave propagation in a Piezo-fiber-reinforced composite (PFRC) layer overlying a functionally graded substrate. When the external characteristic length is comparable to the characteristic length, surface/interface effects become prominent. The dispersion equation has been derived by analytical method. Numerically the effects of volume fraction and surface/interface parameters are analyzed and presented graphically. The results can be found helpful for the development and production of smart materials.
  • Classification : 74G05, 74M05, 74J15
  • Author(s) :
    • Sudarshan Dhua (Department of Mathematics, School of Sciences, NIT Andhra Pradesh)
    • Subrata Mondal (Department of Mathematics, School of Sciences, NIT Andhra Pradesh)