Resolving Intent Ambiguities by Retrieving Discriminative Clarifying\n Questions

Abstract

Task oriented Dialogue Systems generally employ intent detection systems in\norder to map user queries to a set of pre-defined intents. However, user\nqueries appearing in natural language can be easily ambiguous and hence such a\ndirect mapping might not be straightforward harming intent detection and\neventually the overall performance of a dialogue system. Moreover, acquiring\ndomain-specific clarification questions is costly. In order to disambiguate\nqueries which are ambiguous between two intents, we propose a novel method of\ngenerating discriminative questions using a simple rule based system which can\ntake advantage of any question generation system without requiring annotated\ndata of clarification questions. Our approach aims at discrimination between\ntwo intents but can be easily extended to clarification over multiple intents.\nSeeking clarification from the user to classify user intents not only helps\nunderstand the user intent effectively, but also reduces the roboticity of the\nconversation and makes the interaction considerably natural.\n