Semi-supervised learning with self-supervision for closed and open sets
Licentiatavhandling, 2023
SSL is often studied in the closed-set scenario, where we assume that unlabeled data only contain classes present in the labeled data. More realistically, there is a risk that unlabeled data contain unseen classes, corrupted data, or outliers in other forms. This setting is referred to as open-set semi-supervised learning (OSSL). Many existing methods for OSSL use a procedure that involves selecting samples from unlabeled data that likely belong to the known classes, for inclusion in a traditional SSL objective. This work proposes an alternative approach, SeFOSS, that utilizes all unlabeled data through the inclusion of the self-supervised component proposed by DoubleMatch. Additionally, SeFOSS uses an energy-based method for classifying data as in-distribution (ID) or out-of-distribution (OOD). Experimental evaluation shows that SeFOSS achieves strong results for both closed-set accuracy and OOD detection in many open-set scenarios. Additionally, our results indicate that traditional methods for (closed-set) SSL may perform better in the open-set scenario than what has been previously suggested by other works.
Furthermore, this work proposes another method for OSSL: the Beta-model. This method proposes a novel score for ID/OOD classification and introduces the use of the expectation-maximization algorithm in OSSL, for estimating conditional distributions of scores given ID or OOD data. This method demonstrates state-of-the-art results on many benchmark problems for OSSL.
Semi-supervised learning
classification
deep learning
open-set semi-supervised learning
Författare
Erik Wallin
Chalmers, Elektroteknik, Signalbehandling och medicinsk teknik
DoubleMatch: Improving Semi-Supervised Learning with Self-Supervision
Proceedings - International Conference on Pattern Recognition,;Vol. 2022-August(2022)p. 2871-2877
Paper i proceeding
Improving Open-Set Semi-Supervised Learning with Self-Supervision
Proceedings - 2024 IEEE Winter Conference on Applications of Computer Vision, WACV 2024,;(2024)p. 2345-2354
Paper i proceeding
Wallin, E., Svensson, L., Kahl, F., Hammarstrand, L. Beta-model: Open-Set Semi-Supervised Learning with In-Distribution Subspaces
Robust och precis semi-övervakad inlärning
Wallenberg AI, Autonomous Systems and Software Program, 2020-08-25 -- 2024-08-23.
Infrastruktur
C3SE (Chalmers Centre for Computational Science and Engineering)
Ämneskategorier
Signalbehandling
Datavetenskap (datalogi)
Datorseende och robotik (autonoma system)
Utgivare
Chalmers