Your data is not perfect: Towards cross-domain out-of-distribution detection in class-imbalanced data

Abstract

Out-of-distribution detection (OOD detection) aims to detect test samples drawn from a distribution that is different from the training distribution, in order to prevent models trained on in-distribution (ID) data from providing unavailable outputs. Current OOD detection systems typically refer to a single-domain class-balanced assumption that both the training and testing sets belong to the same domain and each class has the same size. Unfortunately, most real-world datasets contain multiple domains and class-imbalanced distributions, which severely limits the applicability of existing works. Previous OOD detection systems only focus on the semantic gap between ID and OOD samples. Besides the semantic gap, we are faced with two additional gaps: the domain gap between source and target domains, and the class-imbalance gap between different classes. In fact, similar objects from different domains should belong to the same class. In this paper, we introduce a realistic yet challenging setting: class-imbalanced cross-domain OOD detection (CCOD), which contains a well-labeled (but usually small) source set for training and conducts OOD detection on an unlabeled (but usually larger) target set for testing. We do not assume that the target domain contains only OOD classes or that it is class-balanced: the distribution among classes of the target dataset need not be the same as the source dataset. To tackle this challenging setting with an OOD detection system, we propose a novel uncertainty-aware adaptive semantic alignment (UASA) network based on a prototype-based alignment strategy. Specifically, we first build label-driven prototypes in the source domain and utilize these prototypes for target classification to close the domain gap. Rather than utilizing fixed thresholds for OOD detection, we generate adaptive sample-wise thresholds to handle the semantic gap. Finally, we conduct uncertainty-aware clustering to group semantically similar target samples to relieve the class-imbalance gap. Extensive experiments on three challenging benchmarks (Office-Home, VisDA-C and DomainNet) demonstrate that our proposed UASA outperforms state-of-the-art methods by a large margin.