DAT: Dynamic Alpha Tuning for Hybrid Retrieval in Retrieval-Augmented Generation

TL;DR

DAT introduces a dynamic, query-aware hybrid retrieval method that leverages large language models to adaptively balance dense and sparse retrieval techniques, significantly improving retrieval effectiveness in RAG systems.

Contribution

The paper presents DAT, a novel framework that dynamically tunes the weighting between dense and sparse retrieval methods using LLM-based effectiveness evaluation, enhancing adaptability and performance.

Findings

DAT outperforms fixed-weight hybrid retrieval methods across multiple metrics.

Even smaller models with DAT achieve strong retrieval performance.

DAT demonstrates efficiency and adaptability in various retrieval scenarios.

Abstract

Hybrid retrieval techniques in Retrieval-Augmented Generation (RAG) systems enhance information retrieval by combining dense and sparse (e.g., BM25-based) retrieval methods. However, existing approaches struggle with adaptability, as fixed weighting schemes fail to adjust to different queries. To address this, we propose DAT (Dynamic Alpha Tuning), a novel hybrid retrieval framework that dynamically balances dense retrieval and BM25 for each query. DAT leverages a large language model (LLM) to evaluate the effectiveness of the top-1 results from both retrieval methods, assigning an effectiveness score to each. It then calibrates the optimal weighting factor through effectiveness score normalization, ensuring a more adaptive and query-aware weighting between the two approaches. Empirical results show that DAT consistently significantly outperforms fixed-weighting hybrid retrieval methods across various evaluation metrics. Even on smaller models, DAT delivers strong performance, highlighting its efficiency and adaptability.