A Unified Approach to Entity-Centric Context Tracking in Social Conversations
This work addresses the challenge of tracking entities and their properties in conversations, which is important for improving conversational AI systems, but it appears incremental as it builds on existing subtasks.
The paper tackles the problem of entity-centric context tracking in social conversations by proposing an end-to-end modeling approach with an entity repository, and it releases a large-scale corpus (Contrack) and a neural network architecture, comparing it to state-of-the-art methods for subtasks.
In human-human conversations, Context Tracking deals with identifying important entities and keeping track of their properties and relationships. This is a challenging problem that encompasses several subtasks such as slot tagging, coreference resolution, resolving plural mentions and entity linking. We approach this problem as an end-to-end modeling task where the conversational context is represented by an entity repository containing the entity references mentioned so far, their properties and the relationships between them. The repository is updated turn-by-turn, thus making training and inference computationally efficient even for long conversations. This paper lays the groundwork for an investigation of this framework in two ways. First, we release Contrack, a large scale human-human conversation corpus for context tracking with people and location annotations. It contains over 7000 conversations with an average of 11.8 turns, 5.8 entities and 15.2 references per conversation. Second, we open-source a neural network architecture for context tracking. Finally we compare this network to state-of-the-art approaches for the subtasks it subsumes and report results on the involved tradeoffs.