English

Selective classification using a robust meta-learning approach

Machine Learning 2024-01-04 v2

Abstract

Predictive uncertainty-a model's self awareness regarding its accuracy on an input-is key for both building robust models via training interventions and for test-time applications such as selective classification. We propose a novel instance-conditioned reweighting approach that captures predictive uncertainty using an auxiliary network and unifies these train- and test-time applications. The auxiliary network is trained using a meta-objective in a bilevel optimization framework. A key contribution of our proposal is the meta-objective of minimizing the dropout variance, an approximation of Bayesian Predictive uncertainty. We show in controlled experiments that we effectively capture the diverse specific notions of uncertainty through this meta-objective, while previous approaches only capture certain aspects. These results translate to significant gains in real-world settings-selective classification, label noise, domain adaptation, calibration-and across datasets-Imagenet, Cifar100, diabetic retinopathy, Camelyon, WILDs, Imagenet-C,-A,-R, Clothing1M, etc. For Diabetic Retinopathy, we see upto 3.4%/3.3% accuracy and AUC gains over SOTA in selective classification. We also improve upon large-scale pretrained models such as PLEX.

Keywords

Cite

@article{arxiv.2212.05987,
  title  = {Selective classification using a robust meta-learning approach},
  author = {Nishant Jain and Karthikeyan Shanmugam and Pradeep Shenoy},
  journal= {arXiv preprint arXiv:2212.05987},
  year   = {2024}
}
R2 v1 2026-06-28T07:31:15.290Z