Conformal Credal Self-Supervised Learning

TL;DR

This paper introduces a theoretically grounded conformal prediction approach to credal self-supervised learning, improving the calibration and reliability of pseudo-labels in semi-supervised learning tasks.

Contribution

It provides a rigorous theoretical foundation for credal sets in self-supervised learning using conformal prediction, along with effective algorithms and empirical validation.

Findings

Enhanced calibration of pseudo-supervision

Competitive performance on benchmark datasets

Theoretically justified uncertainty-aware labeling

Abstract

In semi-supervised learning, the paradigm of self-training refers to the idea of learning from pseudo-labels suggested by the learner itself. Across various domains, corresponding methods have proven effective and achieve state-of-the-art performance. However, pseudo-labels typically stem from ad-hoc heuristics, relying on the quality of the predictions though without guaranteeing their validity. One such method, so-called credal self-supervised learning, maintains pseudo-supervision in the form of sets of (instead of single) probability distributions over labels, thereby allowing for a flexible yet uncertainty-aware labeling. Again, however, there is no justification beyond empirical effectiveness. To address this deficiency, we make use of conformal prediction, an approach that comes with guarantees on the validity of set-valued predictions. As a result, the construction of credal sets of labels is supported by a rigorous theoretical foundation, leading to better calibrated and less error-prone supervision for unlabeled data. Along with this, we present effective algorithms for learning from credal self-supervision. An empirical study demonstrates excellent calibration properties of the pseudo-supervision, as well as the competitiveness of our method on several benchmark datasets.