Converse, Focus and Guess -- Towards Multi-Document Driven Dialogue
This work addresses the problem of efficiently identifying a user's target document from a large collection through dialogue, which is significant for improving information retrieval and conversational AI systems.
This paper introduces Multi-Document Driven Dialogue (MD3), a new task where an agent converses with a user to guess a target document from a large set. They created GuessMovie, a dataset of 16,881 movie documents and 13,434 dialogues, and developed an MD3 model that significantly outperforms baselines and approaches human performance.
We propose a novel task, Multi-Document Driven Dialogue (MD3), in which an agent can guess the target document that the user is interested in by leading a dialogue. To benchmark progress, we introduce a new dataset of GuessMovie, which contains 16,881 documents, each describing a movie, and associated 13,434 dialogues. Further, we propose the MD3 model. Keeping guessing the target document in mind, it converses with the user conditioned on both document engagement and user feedback. In order to incorporate large-scale external documents into the dialogue, it pretrains a document representation which is sensitive to attributes it talks about an object. Then it tracks dialogue state by detecting evolvement of document belief and attribute belief, and finally optimizes dialogue policy in principle of entropy decreasing and reward increasing, which is expected to successfully guess the user's target in a minimum number of turns. Experiments show that our method significantly outperforms several strong baseline methods and is very close to human's performance.