ProSub: Probabilistic Open-Set Semi-Supervised Learning with Subspace-Based Out-of-Distribution Detection

TL;DR

ProSub introduces a probabilistic framework for open-set semi-supervised learning that leverages subspace-based out-of-distribution detection, achieving state-of-the-art results by modeling score distributions and using angular features.

Contribution

It proposes a novel angle-based score for ID/OOD detection and a probabilistic approach to estimate score distributions, enhancing open-set semi-supervised learning.

Findings

Achieves state-of-the-art performance on benchmark datasets.

Introduces a probabilistic model for ID/OOD classification.

Utilizes subspace angles for improved out-of-distribution detection.

Abstract

In open-set semi-supervised learning (OSSL), we consider unlabeled datasets that may contain unknown classes. Existing OSSL methods often use the softmax confidence for classifying data as in-distribution (ID) or out-of-distribution (OOD). Additionally, many works for OSSL rely on ad-hoc thresholds for ID/OOD classification, without considering the statistics of the problem. We propose a new score for ID/OOD classification based on angles in feature space between data and an ID subspace. Moreover, we propose an approach to estimate the conditional distributions of scores given ID or OOD data, enabling probabilistic predictions of data being ID or OOD. These components are put together in a framework for OSSL, termed ProSub, that is experimentally shown to reach SOTA performance on several benchmark problems. Our code is available at https://github.com/walline/prosub.