Zero-Shot Retrieval with Search Agents and Hybrid Environments

TL;DR

This paper introduces a hybrid search environment and agents that improve zero-shot retrieval performance by combining learned query reformulation with heuristic strategies, outperforming traditional neural retrievers.

Contribution

It extends learning to search with hybrid environments, demonstrating that search agents trained with behavioral cloning can outperform baseline retrieval systems and match state-of-the-art results.

Findings

Search agents outperform baseline retrieval systems.

Heuristic Hybrid Retrieval Environments improve performance by several nDCG points.

HARE matches state-of-the-art performance with interpretable actions and higher speed.

Abstract

Learning to search is the task of building artificial agents that learn to autonomously use a search box to find information. So far, it has been shown that current language models can learn symbolic query reformulation policies, in combination with traditional term-based retrieval, but fall short of outperforming neural retrievers. We extend the previous learning to search setup to a hybrid environment, which accepts discrete query refinement operations, after a first-pass retrieval step via a dual encoder. Experiments on the BEIR task show that search agents, trained via behavioral cloning, outperform the underlying search system based on a combined dual encoder retriever and cross encoder reranker. Furthermore, we find that simple heuristic Hybrid Retrieval Environments (HRE) can improve baseline performance by several nDCG points. The search agent based on HRE (HARE) matches state-of-the-art performance, balanced in both zero-shot and in-domain evaluations, via interpretable actions, and at twice the speed.