Semi-Supervised Learning guided by the Generalized Bayes Rule under Soft Revision

Dietrich, Stefan and Rodemann, Julian and Jansen, Christoph (2024) Semi-Supervised Learning guided by the Generalized Bayes Rule under Soft Revision. Other. Arxiv.

[thumbnail of 2405.15294v2]
Text (2405.15294v2)
2405.15294v2.pdf - Published Version
Available under License Creative Commons Attribution.

Download (456kB)

Abstract

We provide a theoretical and computational investigation of the Gamma-Maximin method with soft revision, which was recently proposed as a robust criterion for pseudo-label selection (PLS) in semi-supervised learning. Opposed to traditional methods for PLS we use credal sets of priors ("generalized Bayes") to represent the epistemic modeling uncertainty. These latter are then updated by the Gamma-Maximin method with soft revision. We eventually select pseudo-labeled data that are most likely in light of the least favorable distribution from the so updated credal set. We formalize the task of finding optimal pseudo-labeled data w.r.t. the Gamma-Maximin method with soft revision as an optimization problem. A concrete implementation for the class of logistic models then allows us to compare the predictive power of the method with competing approaches. It is observed that the Gamma-Maximin method with soft revision can achieve very promising results, especially when the proportion of labeled data is low.

Item Type:
Monograph (Other)
Additional Information:
Accepted at the 11th International Conference on Soft Methods in Probability and Statistics (SMPS) 2024
Subjects:
?? stat.mlcs.aics.lgmath.ststat.mestat.th62c12 62c10i.2.6; g.3 ??
ID Code:
221180
Deposited By:
Deposited On:
07 Jun 2024 10:50
Refereed?:
No
Published?:
Published
Last Modified:
14 Jun 2024 23:40