Semi-Supervised Learning with Variational Bayesian Inference and Maximum Uncertainty Regularization

TL;DR

This paper introduces two novel semi-supervised learning methods: one integrating variational Bayesian inference with weight perturbation, and another using maximum uncertainty regularization to enhance model robustness and accuracy.

Contribution

It presents a new approach combining variational Bayesian inference with consistency regularization and introduces maximum uncertainty regularization for improved semi-supervised learning.

Findings

Improved classification accuracy with combined methods

VBI enhances robustness of consistency regularization

MUR broadens smoothness in input-output space

Abstract

We propose two generic methods for improving semi-supervised learning (SSL). The first integrates weight perturbation (WP) into existing "consistency regularization" (CR) based methods. We implement WP by leveraging variational Bayesian inference (VBI). The second method proposes a novel consistency loss called "maximum uncertainty regularization" (MUR). While most consistency losses act on perturbations in the vicinity of each data point, MUR actively searches for "virtual" points situated beyond this region that cause the most uncertain class predictions. This allows MUR to impose smoothness on a wider area in the input-output manifold. Our experiments show clear improvements in classification errors of various CR based methods when they are combined with VBI or MUR or both.