UDAPDR: Unsupervised Domain Adaptation via LLM Prompting and Distillation of Rerankers

TL;DR

This paper introduces UDAPDR, a method that leverages large language models to generate synthetic queries for domain adaptation in information retrieval, improving zero-shot accuracy and reducing latency.

Contribution

It proposes a novel approach combining LLM prompting and distillation of rerankers to enhance domain adaptation without labeled data.

Findings

Boosts zero-shot accuracy in long-tail domains

Achieves lower latency than standard reranking methods

Effective in domain adaptation for IR tasks

Abstract

Many information retrieval tasks require large labeled datasets for fine-tuning. However, such datasets are often unavailable, and their utility for real-world applications can diminish quickly due to domain shifts. To address this challenge, we develop and motivate a method for using large language models (LLMs) to generate large numbers of synthetic queries cheaply. The method begins by generating a small number of synthetic queries using an expensive LLM. After that, a much less expensive one is used to create large numbers of synthetic queries, which are used to fine-tune a family of reranker models. These rerankers are then distilled into a single efficient retriever for use in the target domain. We show that this technique boosts zero-shot accuracy in long-tail domains and achieves substantially lower latency than standard reranking methods.