Timezone: »

 
Poster
Training Over-parameterized Models with Non-decomposable Objectives
Harikrishna Narasimhan · Aditya Menon

Thu Dec 09 08:30 AM -- 10:00 AM (PST) @
in Poster Session 6 »

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.

Keywords: Optimization · Machine Learning · Fairness

Author Information

Harikrishna Narasimhan (Google Research)
Aditya Menon (Google)

More from the Same Authors