Semantically Coherent Out-of-Distribution Detection

TL;DR

This paper introduces the SC-OOD benchmarks to evaluate out-of-distribution detection more realistically and proposes an unsupervised dual grouping framework leveraging unlabeled data to improve detection and classification.

Contribution

It redefines OOD detection benchmarks for semantic coherence and develops a novel unsupervised dual grouping method that enhances semantic understanding and detection performance.

Findings

Existing methods perform poorly on SC-OOD benchmarks.

The proposed UDG framework improves OOD detection accuracy.

State-of-the-art results achieved on new benchmarks.

Abstract

Current out-of-distribution (OOD) detection benchmarks are commonly built by defining one dataset as in-distribution (ID) and all others as OOD. However, these benchmarks unfortunately introduce some unwanted and impractical goals, e.g., to perfectly distinguish CIFAR dogs from ImageNet dogs, even though they have the same semantics and negligible covariate shifts. These unrealistic goals will result in an extremely narrow range of model capabilities, greatly limiting their use in real applications. To overcome these drawbacks, we re-design the benchmarks and propose the semantically coherent out-of-distribution detection (SC-OOD). On the SC-OOD benchmarks, existing methods suffer from large performance degradation, suggesting that they are extremely sensitive to low-level discrepancy between data sources while ignoring their inherent semantics. To develop an effective SC-OOD detection approach, we leverage an external unlabeled set and design a concise framework featured by unsupervised dual grouping (UDG) for the joint modeling of ID and OOD data. The proposed UDG can not only enrich the semantic knowledge of the model by exploiting unlabeled data in an unsupervised manner, but also distinguish ID/OOD samples to enhance ID classification and OOD detection tasks simultaneously. Extensive experiments demonstrate that our approach achieves state-of-the-art performance on SC-OOD benchmarks. Code and benchmarks are provided on our project page: https://jingkang50.github.io/projects/scood.