Predict the Retrieval! Test time adaptation for Retrieval Augmented Generation

TL;DR

This paper introduces TTARAG, a test-time adaptation method that dynamically updates retrieval-augmented generation models during inference to improve performance in specialized domains, addressing distribution shift challenges.

Contribution

The paper presents a novel test-time adaptation technique for RAG systems that enables dynamic parameter updates during inference for domain-specific improvements.

Findings

Significant performance gains over baseline RAG models in six specialized domains.

Effective automatic parameter adjustment through predicting retrieved content.

Demonstrated robustness across diverse domain shifts.

Abstract

Retrieval-Augmented Generation (RAG) has emerged as a powerful approach for enhancing large language models' question-answering capabilities through the integration of external knowledge. However, when adapting RAG systems to specialized domains, challenges arise from distribution shifts, resulting in suboptimal generalization performance. In this work, we propose TTARAG, a test-time adaptation method that dynamically updates the language model's parameters during inference to improve RAG system performance in specialized domains. Our method introduces a simple yet effective approach where the model learns to predict retrieved content, enabling automatic parameter adjustment to the target domain. Through extensive experiments across six specialized domains, we demonstrate that TTARAG achieves substantial performance improvements over baseline RAG systems. Code available at https://github.com/sunxin000/TTARAG.