ConvSDG: Session Data Generation for Conversational Search

TL;DR

ConvSDG leverages large language models to generate synthetic conversational session data, enhancing training for conversational dense retrieval and improving search effectiveness across multiple datasets.

Contribution

Proposes a novel framework using LLMs for session data generation to improve conversational search models, addressing data scarcity issues.

Findings

Generated data improves retrieval performance

Effective across multiple datasets

Outperforms several strong baselines

Abstract

Conversational search provides a more convenient interface for users to search by allowing multi-turn interaction with the search engine. However, the effectiveness of the conversational dense retrieval methods is limited by the scarcity of training data required for their fine-tuning. Thus, generating more training conversational sessions with relevant labels could potentially improve search performance. Based on the promising capabilities of large language models (LLMs) on text generation, we propose ConvSDG, a simple yet effective framework to explore the feasibility of boosting conversational search by using LLM for session data generation. Within this framework, we design dialogue/session-level and query-level data generation with unsupervised and semi-supervised learning, according to the availability of relevance judgments. The generated data are used to fine-tune the conversational dense retriever. Extensive experiments on four widely used datasets demonstrate the effectiveness and broad applicability of our ConvSDG framework compared with several strong baselines.