On the Powerfulness of Textual Outlier Exposure for Visual OoD Detection

TL;DR

This paper explores the use of textual outlier exposure to improve visual out-of-distribution detection, demonstrating that generated textual outliers can enhance performance and provide insights into designing effective outliers.

Contribution

It introduces textual outlier exposure for OoD detection, expanding beyond visual outliers, and provides criteria for designing effective textual outliers based on empirical analysis.

Findings

Textual outliers improve OoD detection performance.

Generated textual outliers are competitive on large-scale benchmarks.

Design criteria include near-distribution, descriptiveness, and visual semantics inclusion.

Abstract

Successful detection of Out-of-Distribution (OoD) data is becoming increasingly important to ensure safe deployment of neural networks. One of the main challenges in OoD detection is that neural networks output overconfident predictions on OoD data, make it difficult to determine OoD-ness of data solely based on their predictions. Outlier exposure addresses this issue by introducing an additional loss that encourages low-confidence predictions on OoD data during training. While outlier exposure has shown promising potential in improving OoD detection performance, all previous studies on outlier exposure have been limited to utilizing visual outliers. Drawing inspiration from the recent advancements in vision-language pre-training, this paper venture out to the uncharted territory of textual outlier exposure. First, we uncover the benefits of using textual outliers by replacing real or virtual outliers in the image-domain with textual equivalents. Then, we propose various ways of generating preferable textual outliers. Our extensive experiments demonstrate that generated textual outliers achieve competitive performance on large-scale OoD and hard OoD benchmarks. Furthermore, we conduct empirical analyses of textual outliers to provide primary criteria for designing advantageous textual outliers: near-distribution, descriptiveness, and inclusion of visual semantics.