Towards Deep Conversational Recommendations

Citations Over Time

Abstract

There has been growing interest in using neural networks and deep learning techniques to create dialogue systems. Conversational recommendation is an interesting setting for the scientific exploration of dialogue with natural language as the associated discourse involves goal-driven dialogue that often transforms naturally into more free-form chat. This paper provides two contributions. First, until now there has been no publicly available large-scale dataset consisting of real-world dialogues centered around recommendations. To address this issue and to facilitate our exploration here, we have collected ReDial, a dataset consisting of over 10,000 conversations centered around the theme of providing movie recommendations. We make this data available to the community for further research. Second, we use this dataset to explore multiple facets of conversational recommendations. In particular we explore new neural architectures, mechanisms, and methods suitable for composing conversational recommendation systems. Our dataset allows us to systematically probe model sub-components addressing different parts of the overall problem domain ranging from: sentiment analysis and cold-start recommendation generation to detailed aspects of how natural language is used in this setting in the real world. We combine such sub-components into a full-blown dialogue system and examine its behavior.

Related Papers

• → (2019)31,543 cited
• → Towards Conversational Recommender Systems(2016)400 cited
• → Conversational Recommender System(2018)363 cited
• → Towards Conversational Search and Recommendation(2018)323 cited
• → Estimation-Action-Reflection: Towards Deep Interaction Between Conversational and Recommender Systems(2020)243 cited