Active Learning for Matching Problems

TL;DR

This paper introduces a new active learning approach tailored for matching problems, focusing on reducing user input by selecting informative queries based on probabilistic matchings and matching objectives.

Contribution

It presents a novel probabilistic matching method and develops active learning strategies specifically designed for matching objectives, improving preference learning efficiency.

Findings

Matching-sensitive active learning outperforms baseline methods

Significant reduction in user preference queries needed

Effective across diverse real-world datasets

Abstract

Effective learning of user preferences is critical to easing user burden in various types of matching problems. Equally important is active query selection to further reduce the amount of preference information users must provide. We address the problem of active learning of user preferences for matching problems, introducing a novel method for determining probabilistic matchings, and developing several new active learning strategies that are sensitive to the specific matching objective. Experiments with real-world data sets spanning diverse domains demonstrate that matching-sensitive active learning