无需标注数据，用大模型自动生成澄清需求预测训练数据。
Zero-Shot and Efficient Clarification Need Prediction in Conversational Search
- 用大模型零样本生成模糊与明确查询对，替代人工标注。
- 在多个数据集上性能超越现有零样本与少样本方法。
- 适合需要快速部署且无标注数据的对话搜索场景。
澄清需求预测（CNP）是对话搜索中的关键任务，旨在判断是否应提出澄清问题或直接回答用户查询。当前研究受限于标注数据稀缺和推理效率低。本文提出零样本高效CNP框架Zef-CNP：首先通过零样本提示让大语言模型（LLM）生成两类合成查询——模糊与明确查询；再利用这些生成数据训练轻量级CNP模型。该方法避免了训练阶段的人工标注，也无需在查询时使用高延迟的LLM。为提升生成质量，设计了面向主题、信息需求与查询的思维链提示（TIQ-CoT），并引入反事实查询生成（CoQu）策略，引导模型先生成特定/模糊查询，再生成对应的模糊/特定版本。实验表明，Zef-CNP在有效性和效率上均优于零样本与少样本基于LLM的预测器。
原文摘要 · Abstract (English)
Clarification need prediction (CNP) is a key task in conversational search, aiming to predict whether to ask a clarifying question or give an answer to the current user query. However, current research on CNP suffers from the issues of limited CNP training data and low efficiency. In this paper, we propose a zero-shot and efficient CNP framework (Zef-CNP), in which we first prompt large language models (LLMs) in a zero-shot manner to generate two sets of synthetic queries: ambiguous and specific (unambiguous) queries. We then use the generated queries to train efficient CNP models. Zef-CNP eliminates the need for human-annotated clarification-need labels during training and avoids the use of LLMs with high query latency at query time. To further improve the generation quality of synthetic queries, we devise a topic-, information-need-, and query-aware chain-of-thought (CoT) prompting strategy (TIQ-CoT). Moreover, we enhance TIQ-CoT with counterfactual query generation (CoQu), which guides LLMs first to generate a specific/ambiguous query and then sequentially generate its corresponding ambiguous/specific query. Experimental results show that Zef-CNP achieves superior CNP effectiveness and efficiency compared with zero- and few-shot LLM-based CNP predictors.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。