Query refinement using conversational context: A method and an evaluation resource

Abstract

This paper introduces a query refinement method applied to queries asked by users during a meeting or a conversation. The proposed method does not require further clarifications from users, to avoid distracting them from their conversation, but leverages instead the local context of the conversation. The method first represents the local context by extracting keywords from the transcript of the conversation. It then expands the queries with keywords that best represent the topic of the query, i.e. expansion keywords accompanied by weights indicating their topical similarity to the query. Moreover, we present a dataset called AREX and an evaluation metric based on relevance judgments collected in a crowdsourcing experiment. We compare our query expansion approach with other methods, over queries extracted from the AREX dataset, showing the superiority of our method when either manual or automatic transcripts of the AMI Meeting Corpus are used.