Session Time & Room : 4E (Aug.24, 17:40-19:20) @E711
Type : Contributed Talk
Abstract : The empirical risk minimization (ERM) approach for supervised learning has been the workhorse of machine learning. However, ERM methods strongly rely on the specific training samples available and cannot easily address scenarios affected by distribution shifts and corrupted samples. This talk presents a learning framework based on the generalized maximum entropy principle that leads to minimax risk classifiers (MRCs). MRC learning is based on expectation estimates and does not strongly rely on specific training samples.
