Learning When to Say "I Don't Know"

TL;DR

This paper introduces a novel reject option classification method that uses per-class softmax thresholds to improve decision accuracy by effectively identifying uncertain regions, outperforming traditional thresholding approaches across various datasets.

Contribution

The paper presents a new approach for reject option classification that leverages validation set analysis to learn softmax thresholds, avoiding the need for known rejection costs or strict accuracy constraints.

Findings

Outperforms naive softmax thresholding on multiple datasets

Effective in image, text, and 2-D classification tasks

Utilizes validation set to optimize rejection thresholds

Abstract

We propose a new Reject Option Classification technique to identify and remove regions of uncertainty in the decision space for a given neural classifier and dataset. Such existing formulations employ a learned rejection (remove)/selection (keep) function and require either a known cost for rejecting examples or strong constraints on the accuracy or coverage of the selected examples. We consider an alternative formulation by instead analyzing the complementary reject region and employing a validation set to learn per-class softmax thresholds. The goal is to maximize the accuracy of the selected examples subject to a natural randomness allowance on the rejected examples (rejecting more incorrect than correct predictions). We provide results showing the benefits of the proposed method over na\"ively thresholding calibrated/uncalibrated softmax scores with 2-D points, imagery, and text classification datasets using state-of-the-art pretrained models. Source code is available at https://github.com/osu-cvl/learning-idk.