PAC Prediction Sets for Meta-Learning

TL;DR

This paper introduces a novel PAC prediction set method for meta-learning that provides reliable uncertainty quantification across new tasks with minimal data, demonstrating improved set size and guarantee satisfaction.

Contribution

The paper proposes a new algorithm for PAC prediction sets in meta-learning, extending the guarantee to the meta setting and showing effectiveness across diverse domains.

Findings

Prediction sets satisfy PAC guarantees across tasks

Smaller set sizes compared to baselines

Effective in visual, language, and medical domains

Abstract

Uncertainty quantification is a key component of machine learning models targeted at safety-critical systems such as in healthcare or autonomous vehicles. We study this problem in the context of meta learning, where the goal is to quickly adapt a predictor to new tasks. In particular, we propose a novel algorithm to construct \emph{PAC prediction sets}, which capture uncertainty via sets of labels, that can be adapted to new tasks with only a few training examples. These prediction sets satisfy an extension of the typical PAC guarantee to the meta learning setting; in particular, the PAC guarantee holds with high probability over future tasks. We demonstrate the efficacy of our approach on four datasets across three application domains: mini-ImageNet and CIFAR10-C in the visual domain, FewRel in the language domain, and the CDC Heart Dataset in the medical domain. In particular, our prediction sets satisfy the PAC guarantee while having smaller size compared to other baselines that also satisfy this guarantee.