Feasible Learning
This addresses the issue of poor performance on outliers or difficult samples for practitioners in fields like computer vision and natural language processing, representing an incremental advance over Empirical Risk Minimization.
The authors tackled the problem of improving worst-case performance in machine learning by introducing Feasible Learning (FL), a paradigm that bounds loss for each training sample, and demonstrated that FL models show improved tail behavior with only a marginal impact on average performance in tasks like image classification and age regression.
We introduce Feasible Learning (FL), a sample-centric learning paradigm where models are trained by solving a feasibility problem that bounds the loss for each training sample. In contrast to the ubiquitous Empirical Risk Minimization (ERM) framework, which optimizes for average performance, FL demands satisfactory performance on every individual data point. Since any model that meets the prescribed performance threshold is a valid FL solution, the choice of optimization algorithm and its dynamics play a crucial role in shaping the properties of the resulting solutions. In particular, we study a primal-dual approach which dynamically re-weights the importance of each sample during training. To address the challenge of setting a meaningful threshold in practice, we introduce a relaxation of FL that incorporates slack variables of minimal norm. Our empirical analysis, spanning image classification, age regression, and preference optimization in large language models, demonstrates that models trained via FL can learn from data while displaying improved tail behavior compared to ERM, with only a marginal impact on average performance.