Mind the Gap Between Conversations for Improved Long-Term Dialogue Generation
This work addresses the challenge of making dialogue systems more realistic for long-term human-computer interactions, though it is incremental as it builds on existing dialogue modeling with a new dataset and time representations.
The paper tackled the problem of dialogue systems generating unnatural responses by not accounting for time gaps between conversations, and introduced GapChat, a multi-session dataset with simulated event progress to model time-aware dialogues, showing in human evaluation that time-aware models improved topic relevance and information gain.
Knowing how to end and resume conversations over time is a natural part of communication, allowing for discussions to span weeks, months, or years. The duration of gaps between conversations dictates which topics are relevant and which questions to ask, and dialogue systems which do not explicitly model time may generate responses that are unnatural. In this work we explore the idea of making dialogue models aware of time, and present GapChat, a multi-session dialogue dataset in which the time between each session varies. While the dataset is constructed in real-time, progress on events in speakers' lives is simulated in order to create realistic dialogues occurring across a long timespan. We expose time information to the model and compare different representations of time and event progress. In human evaluation we show that time-aware models perform better in metrics that judge the relevance of the chosen topics and the information gained from the conversation.