Topological Structure Learning for Weakly-Supervised Out-of-Distribution Detection

TL;DR

This paper introduces a novel weakly-supervised setting for out-of-distribution detection, leveraging limited labeled data and abundant unlabeled data, and proposes a topological structure learning method to improve detection accuracy.

Contribution

The paper proposes a new WSOOD setting and introduces TSL, a topological structure learning approach that enhances ID and OOD separation with limited labeled data.

Findings

TSL significantly outperforms existing methods on multiple datasets.

The method effectively utilizes limited labeled data and unlabeled data.

TSL demonstrates robustness and validity in the new WSOOD setting.

Abstract

Out-of-distribution (OOD) detection is the key to deploying models safely in the open world. For OOD detection, collecting sufficient in-distribution (ID) labeled data is usually more time-consuming and costly than unlabeled data. When ID labeled data is limited, the previous OOD detection methods are no longer superior due to their high dependence on the amount of ID labeled data. Based on limited ID labeled data and sufficient unlabeled data, we define a new setting called Weakly-Supervised Out-of-Distribution Detection (WSOOD). To solve the new problem, we propose an effective method called Topological Structure Learning (TSL). Firstly, TSL uses a contrastive learning method to build the initial topological structure space for ID and OOD data. Secondly, TSL mines effective topological connections in the initial topological space. Finally, based on limited ID labeled data and mined topological connections, TSL reconstructs the topological structure in a new topological space to increase the separability of ID and OOD instances. Extensive studies on several representative datasets show that TSL remarkably outperforms the state-of-the-art, verifying the validity and robustness of our method in the new setting of WSOOD.