Making Better Use of Unlabelled Data in Bayesian Active Learning

Freddie Bickford Smith, Adam Foster, Tom Rainforth
4/26/2024
cs.LGstat.ML

No official code found. Best implementation: fbickfordsmith/epig (41 stars, MIT, last commit 2025-09-04).

Abstract

Fully supervised models are predominant in Bayesian active learning. We argue that their neglect of the information present in unlabelled data harms not just predictive performance but also decisions about what data to acquire. Our proposed solution is a simple framework for semi-supervised Bayesian active learning. We find it produces better-performing models than either conventional Bayesian active learning or semi-supervised learning with randomly acquired data. It is also easier to scale up than the conventional approach. As well as supporting a shift towards semi-supervised models, our findings highlight the importance of studying models and acquisition methods in conjunction.

View on arXivView PDF

Code Implementations(1)

fbickfordsmith/epigCommunity100%
MIT

Bayesian active learning with EPIG data acquisition

418Shell, PythonFeb 20, 20231 years agoMIT

Similar Papers

The Role of Active Learning in Modern Machine Learning

85% match2025

Active Semi-Supervised Learning by Exploring Per-Sample Uncertainty and Consistency

85% match2023

Prediction-Oriented Bayesian Active Learning

84% match1 repo412023

Active Learning Methods for Efficient Data Utilization and Model Performance Enhancement

83% match2025

On the Limitations of Simulating Active Learning

82% match2023

To Label or Not to Label: PALM -- A Predictive Model for Evaluating Sample Efficiency in Active Learning Models

82% match1 repo62025

Cite this paper

@article{smith2024making,
  title  = {Making Better Use of Unlabelled Data in Bayesian Active Learning},
  author = {Freddie Bickford Smith and Adam Foster and Tom Rainforth},
  year   = {2024},
  eprint = {2404.17249},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2404.17249v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2404.17249.svg)](https://codeofpaper.com/paper/2404.17249)

Discussion