Adaptive Batch Sizes for Active Learning A Probabilistic Numerics Approach

TL;DR

This paper introduces a probabilistic numerics framework for adaptive batch sizing in active learning, optimizing the trade-off between cost and speed by automatically tuning batch sizes based on integration error, improving efficiency in Bayesian learning and optimization.

Contribution

The paper presents a novel probabilistic numerics approach that adaptively adjusts batch sizes during active learning, eliminating the need for exhaustive batch size searches and enhancing efficiency.

Findings

Significantly improves learning efficiency in active learning.

Automatically adapts batch sizes based on integration error.

Enhances flexibility in Bayesian optimization applications.

Abstract

Active learning parallelization is widely used, but typically relies on fixing the batch size throughout experimentation. This fixed approach is inefficient because of a dynamic trade-off between cost and speed -- larger batches are more costly, smaller batches lead to slower wall-clock run-times -- and the trade-off may change over the run (larger batches are often preferable earlier). To address this trade-off, we propose a novel Probabilistic Numerics framework that adaptively changes batch sizes. By framing batch selection as a quadrature task, our integration-error-aware algorithm facilitates the automatic tuning of batch sizes to meet predefined quadrature precision objectives, akin to how typical optimizers terminate based on convergence thresholds. This approach obviates the necessity for exhaustive searches across all potential batch sizes. We also extend this to scenarios with constrained active learning and constrained optimization, interpreting constraint violations as reductions in the precision requirement, to subsequently adapt batch construction. Through extensive experiments, we demonstrate that our approach significantly enhances learning efficiency and flexibility in diverse Bayesian batch active learning and Bayesian optimization applications.