Making Better Use of Unlabelled Data in Bayesian Active Learning

TL;DR

This paper introduces a semi-supervised Bayesian active learning framework that leverages unlabelled data to improve model performance and data acquisition decisions, outperforming traditional methods and being more scalable.

Contribution

It presents a simple, scalable semi-supervised approach for Bayesian active learning that effectively utilizes unlabelled data, enhancing model accuracy and acquisition efficiency.

Findings

Better model performance than traditional Bayesian active learning.

More scalable than conventional semi-supervised approaches.

Highlights importance of joint study of models and acquisition methods.

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.