Dense Passage Retrieval in Conversational Search

2025-03-21

AI Summary

Traditional sparse retrieval methods (e.g., BM25) struggle to model semantic dependencies across multi-turn conversational contexts in conversational search. Method: We propose GPT2QR+DPR, a lightweight end-to-end dense retrieval framework that integrates GPT-2–driven query rewriting, explicit dialogue history modeling, and dual-encoder dense passage retrieval (DPR), enhancing contextual relevance during first-stage retrieval. Contribution/Results: This work presents the first systematic evaluation of dense retrieval on the CAsT benchmark for conversational search, demonstrating substantial improvements over BM25 without large-scale fine-tuning. Experiments show consistent gains in retrieval accuracy and robustness, validating the generalization capability of dense representations for modeling multi-turn conversational semantics. Our approach establishes a new paradigm for efficient, low-overhead conversational search—achieving strong performance with minimal computational cost and no task-specific pretraining.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: Question Answering
• Search and Optimization: Distributed Search

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Information retrieval systems have traditionally relied on exact term match methods such as BM25 for first-stage retrieval. However, recent advancements in neural network-based techniques have introduced a new method called dense retrieval. This approach uses a dual-encoder to create contextual embeddings that can be indexed and clustered efficiently at run-time, resulting in improved retrieval performance in Open-domain Question Answering systems. In this paper, we apply the dense retrieval technique to conversational search by conducting experiments on the CAsT benchmark dataset. We also propose an end-to-end conversational search system called GPT2QR+DPR, which incorporates various query reformulation strategies to improve retrieval accuracy. Our findings indicate that dense retrieval outperforms BM25 even without extensive fine-tuning. Our work contributes to the growing body of research on neural-based retrieval methods in conversational search, and highlights the potential of dense retrieval in improving retrieval accuracy in conversational search systems.

Problem

• Improving retrieval performance in conversational search systems
• Applying dense retrieval to conversational search using CAsT dataset
• Comparing dense retrieval with traditional BM25 methods

Innovation

• Dual-encoder for contextual embeddings
• GPT2QR+DPR end-to-end system
• Dense retrieval outperforms BM25