Harnessing Evolution of Multi-Turn Conversations for Effective Answer Retrieval

With the improvements in speech recognition and voice generation technologies over the last years, lots of companies have sought to develop conversation understanding systems that run on mobile phones or smart home devices through natural language interfaces. Conversational assistants, such as Google Assistant and Microsoft Cortana, can help users to complete various types of tasks. This requires an accurate understanding of the user's information need as the conversation evolves into multiple turns. Finding relevant context in a conversation's history is challenging because of the complexity of natural language and evolution of user's information need. In this work, we present an extensive analysis on language, relevance, dependency of utterances in a multi-turn information-seeking conversation. To this aim, we have annotated relevant utterances in the conversations released by the TREC CaST track. The annotation labels determine which of the previous utterances in a conversation can be used to improve the current one. Furthermore, we propose a neural utterance relevance model based on BERT fine-tuning, outperforming competitive baselines.
We study and compare the performance of multiple retrieval models, utilizing different strategies to incorporate the user's context. The experimental results on both classification and retrieval tasks show that our proposed approach can effectively identify and incorporate the conversation context. We show that processing the current utterance using the predicted relevant utterance leads to 38% relative improvement in terms of nDCG@20. Moreover, we see that the conversation context has the highest impact on the third turn in a conversation. Finally, to foster research in this area we have released the dataset of the annotations.