Training Over-parameterized Models with Non-decomposable Objectives

Abstract

Many modern machine learning applications come with complex and nuanced design goals such as minimizing the worst-case error, satisfying a given precision or recall target, or enforcing group-fairness constraints. Popular techniques for optimizing such non-decomposable objectives reduce the problem into a sequence of cost-sensitive learning tasks, each of which is then solved by re-weighting the training loss with example-specific costs. We point out that the standard approach of re-weighting the loss to incorporate label costs can produce unsatisfactory results when used to train over-parameterized models. As a remedy, we propose new cost-sensitive losses that extend the classical idea of logit adjustment to handle more general cost matrices. Our losses are calibrated, and can be further improved with distilled labels from a teacher model. Through experiments on benchmark image datasets, we showcase the effectiveness of our approach in training ResNet models with common robust and constrained optimization objectives.

References in corpus (11)

• Distilling the Knowledge in a Neural Network
• Equality of Opportunity in Supervised Learning
• The Devil is in the Tails: Fine-grained Classification in the Wild
• Class-Balanced Loss Based on Effective Number of Samples
• Optimizing F-measure: A Tale of Two Approaches
• Distribution Alignment: A Unified Framework for Long-tail Visual Recognition
• Addressing the Loss-Metric Mismatch with Adaptive Loss Alignment
• Optimizing Non-decomposable Performance Measures: A Tale of Two Classes
• Online and Stochastic Gradient Methods for Non-decomposable Loss Functions
• Balanced Knowledge Distillation for Long-tailed Learning
• Implicit Rate-Constrained Optimization of Non-decomposable Objectives