Asking Clarifying Questions in Open-Domain Information-Seeking Conversations (SIGIR 2019)
Metadata
| Field | Content | |-------|---------| | Title | Asking Clarifying Questions in Open-Domain Information-Seeking Conversations, SIGIR 2019 | | Source | https://dl.acm.org/doi/abs/10.1145/3331184.3331265 | | Type | Academic paper | | Section | Search Assistance |
Overview
This forum entry covers the SIGIR 2019 paper on asking clarifying questions in open-domain information-seeking conversations. The work tackles a classic problem in conversational search: user queries in open-domain settings are often ambiguous, and rather than returning results for a guessed interpretation, the system can proactively ask a clarifying question to narrow down the user's intent.
Key Contributions
- A crowdsourced collection of search conversations in which participants ask clarifying questions and provide answers, enabling study of mixed-initiative interaction in open-domain search.
- An analysis of how clarifying questions influence user engagement and search behavior in conversational settings.
- A ranking model that selects the most appropriate clarifying question given a query and conversation context, demonstrating improved effectiveness over baselines.
- Related entries in this list:
- Asking Clarification Questions to Handle Ambiguity in Open-Domain QA
- DiAL: Diversity aware listwise ranking for query auto-complete, EMNLP
- Enhancing Discoverability in Enterprise Conversational Systems
- Evaluating auto-complete ranking for diversity and relevance, ECIR
- Evaluation and Continual Improvement for an Enterprise AI Assistant
- Generating Query Recommendations via LLMs (arXiv 2024)
- Clarification is a practical alternative to one-shot retrieval when intent is ambiguous; it shifts burden from result ranking to dialogue strategy.
- In modern LLM-based RAG and agentic search stacks, whether to ask, what to ask, and when to stop are learnable decisions that interact with latency and cost budgets.
- Offline metrics alone are insufficient: user-perceived gains (answer trustworthiness, multi-turn consistency) should drive product evaluation.
- Original paper: Asking Clarifying Questions in Open-Domain Information-Seeking Conversations, SIGIR 2019 — https://dl.acm.org/doi/abs/10.1145/3331184.3331265
Context and Relevance
The paper is an early and influential reference for conversational information retrieval and proactive search assistance. It connects to a broader research thread on ambiguity handling, query suggestion, and dialogue-based retrieval, including later work on generating clarification questions for open-domain QA and evaluation of query auto-complete ranking.