Towards More Trustworthy Deep Code Models by Enabling Out-of-Distribution Detection

TL;DR

This paper introduces two novel SE-specific out-of-distribution detection models for deep code models, enhancing their trustworthiness by effectively identifying data that differs from training distributions, with significant experimental improvements.

Contribution

The paper proposes unsupervised and weakly-supervised OOD detection methods tailored for software engineering tasks, addressing the challenge of distribution shifts in deep code models.

Findings

Proposed methods outperform baselines in four OOD scenarios.

Enhanced OOD detection improves main code understanding tasks.

Models effectively handle distribution shifts in real-world settings.

Abstract

Numerous machine learning (ML) models have been developed, including those for software engineering (SE) tasks, under the assumption that training and testing data come from the same distribution. However, training and testing distributions often differ, as training datasets rarely encompass the entire distribution, while testing distribution tends to shift over time. Hence, when confronted with out-of-distribution (OOD) instances that differ from the training data, a reliable and trustworthy SE ML model must be capable of detecting them to either abstain from making predictions, or potentially forward these OODs to appropriate models handling other categories or tasks. In this paper, we develop two types of SE-specific OOD detection models, unsupervised and weakly-supervised OOD detection for code. The unsupervised OOD detection approach is trained solely on in-distribution samples while the weakly-supervised approach utilizes a tiny number of OOD samples to further enhance the detection performance in various OOD scenarios. Extensive experimental results demonstrate that our proposed methods significantly outperform the baselines in detecting OOD samples from four different scenarios simultaneously and also positively impact a main code understanding task.