Leveraging Importance Weights in Subset Selection

TL;DR

This paper introduces IWeS, a subset selection algorithm that uses importance sampling based on model entropy to improve performance in batch settings, with theoretical support and competitive results.

Contribution

The paper proposes IWeS, a novel importance sampling-based subset selection algorithm applicable to arbitrary models in batch settings, with theoretical analysis and empirical validation.

Findings

Significant performance gains over existing algorithms on seven datasets

Competitive results in active learning scenarios without label information

Theoretical bounds supporting the importance weighting approach

Abstract

We present a subset selection algorithm designed to work with arbitrary model families in a practical batch setting. In such a setting, an algorithm can sample examples one at a time but, in order to limit overhead costs, is only able to update its state (i.e. further train model weights) once a large enough batch of examples is selected. Our algorithm, IWeS, selects examples by importance sampling where the sampling probability assigned to each example is based on the entropy of models trained on previously selected batches. IWeS admits significant performance improvement compared to other subset selection algorithms for seven publicly available datasets. Additionally, it is competitive in an active learning setting, where the label information is not available at selection time. We also provide an initial theoretical analysis to support our importance weighting approach, proving generalization and sampling rate bounds.