Prediction-Oriented Subsampling from Data Streams

TL;DR

This paper proposes a prediction-oriented subsampling method for data streams that reduces uncertainty in specific predictions, demonstrating improved performance over previous techniques with careful model design.

Contribution

It introduces an information-theoretic subsampling approach focused on prediction accuracy, advancing data stream learning methods.

Findings

Outperforms previous information-theoretic subsampling methods

Effective in two widely studied problems

Requires careful model design for strong performance

Abstract

Data is often generated in streams, with new observations arriving over time. A key challenge for learning models from data streams is capturing relevant information while keeping computational costs manageable. We explore intelligent data subsampling for offline learning, and argue for an information-theoretic method centred on reducing uncertainty in downstream predictions of interest. Empirically, we demonstrate that this prediction-oriented approach performs better than a previously proposed information-theoretic technique on two widely studied problems. At the same time, we highlight that reliably achieving strong performance in practice requires careful model design.