No True State-of-the-Art? OOD Detection Methods are Inconsistent across Datasets

Abstract

Out-of-distribution detection is an important component of reliable ML systems. Prior literature has proposed various methods (e.g., MSP (Hendrycks & Gimpel, 2017), ODIN (Liang et al., 2018), Mahalanobis (Lee et al., 2018)), claiming they are state-of-the-art by showing they outperform previous methods on a selected set of in-distribution (ID) and out-of-distribution (OOD) datasets. In this work, we show that none of these methods are inherently better at OOD detection than others on a standardized set of 16 (ID, OOD) pairs. We give possible explanations for these inconsistencies with simple toy datasets where whether one method outperforms another depends on the structure of the ID and OOD datasets in question. Finally, we show that a method outperforming another on a certain (ID, OOD) pair may not do so in a low-data regime. In the low-data regime, we propose a distance-based method, Pairwise OOD detection (POD), which is based on Siamese networks and improves over Mahalanobis by sidestepping the expensive covariance estimation step. Our results suggest that the OOD detection problem may be too broad, and we should consider more specific structures for leverage.

arXiv commentICML Workshop on Uncertainty & Robustness in Deep Learning, 2021

References in corpus (12)

• A Simple Framework for Contrastive Learning of Visual Representations
• A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
• LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop
• A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
• Energy-based Out-of-distribution Detection
• Deep Anomaly Detection with Outlier Exposure
• A Simple Fix to Mahalanobis Distance for Improving Near-OOD Detection
• Consistent Estimators for Learning to Defer to an Expert
• CheXpedition: Investigating Generalization Challenges for Translation of Chest X-Ray Algorithms to the Clinical Setting
• Open Set Medical Diagnosis
• Prototypical Clustering Networks for Dermatological Disease Diagnosis
• CheXseen: Unseen Disease Detection for Deep Learning Interpretation of Chest X-rays

Cited by in corpus (2)

• A Unified Benchmark for the Unknown Detection Capability of Deep Neural Networks
• Toward a Realistic Benchmark for Out-of-Distribution Detection