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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dietrich, Stefan, Rodemann, Julian, Jansen, Christoph
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910471622230016
author Dietrich, Stefan
Rodemann, Julian
Jansen, Christoph
author_facet Dietrich, Stefan
Rodemann, Julian
Jansen, Christoph
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2405_15294
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Semi-Supervised Learning guided by the Generalized Bayes Rule under Soft Revision
Dietrich, Stefan
Rodemann, Julian
Jansen, Christoph
Machine Learning
Artificial Intelligence
Statistics Theory
Methodology
62C12 62C10
I.2.6; G.3
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.
title Semi-Supervised Learning guided by the Generalized Bayes Rule under Soft Revision
topic Machine Learning
Artificial Intelligence
Statistics Theory
Methodology
62C12 62C10
I.2.6; G.3
url https://arxiv.org/abs/2405.15294