CLARINET: Augmenting Language Models to Ask Clarification Questions for Retrieval

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chi, Yizhou, Lin, Jessy, Lin, Kevin, Klein, Dan
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911886723776512
author Chi, Yizhou
Lin, Jessy
Lin, Kevin
Klein, Dan
author_facet Chi, Yizhou
Lin, Jessy
Lin, Kevin
Klein, Dan
contents Users often make ambiguous requests that require clarification. We study the problem of asking clarification questions in an information retrieval setting, where systems often face ambiguous search queries and it is challenging to turn the uncertainty in the retrieval model into a natural language question. We present CLARINET, a system that asks informative clarification questions by choosing questions whose answers would maximize certainty in the correct candidate. Our approach works by augmenting a large language model (LLM) to condition on a retrieval distribution, finetuning end-to-end to generate the question that would have maximized the rank of the true candidate at each turn. When evaluated on a real-world retrieval dataset of users searching for books, our system outperforms traditional heuristics such as information gain on retrieval success by 17% and vanilla-prompted LLMs by 39% relative.
format Preprint
id arxiv_https___arxiv_org_abs_2405_15784
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CLARINET: Augmenting Language Models to Ask Clarification Questions for Retrieval
Chi, Yizhou
Lin, Jessy
Lin, Kevin
Klein, Dan
Information Retrieval
Artificial Intelligence
Computation and Language
Users often make ambiguous requests that require clarification. We study the problem of asking clarification questions in an information retrieval setting, where systems often face ambiguous search queries and it is challenging to turn the uncertainty in the retrieval model into a natural language question. We present CLARINET, a system that asks informative clarification questions by choosing questions whose answers would maximize certainty in the correct candidate. Our approach works by augmenting a large language model (LLM) to condition on a retrieval distribution, finetuning end-to-end to generate the question that would have maximized the rank of the true candidate at each turn. When evaluated on a real-world retrieval dataset of users searching for books, our system outperforms traditional heuristics such as information gain on retrieval success by 17% and vanilla-prompted LLMs by 39% relative.
title CLARINET: Augmenting Language Models to Ask Clarification Questions for Retrieval
topic Information Retrieval
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2405.15784