Conversation Chronicles: Towards Diverse Temporal and Relational Dynamics in Multi-Session Conversations

In the field of natural language processing, open-domain chatbots have emerged as an important research topic. However, a major limitation of existing open-domain chatbot research is its singular focus on short single-session dialogue, neglecting the potential need for understanding contextual information in multiple consecutive sessions that precede an ongoing dialogue. Among the elements that compose the context in multi-session conversation settings, the time intervals between sessions and the relationships between speakers would be particularly important. Despite their importance, current research efforts have not sufficiently addressed these dialogical components. In this paper, we introduce a new 1M multisession dialogue dataset, called CONVERSATION CHRONICLES, for implementing a long-term conversation setup in which time intervals and fine-grained speaker relationships are incorporated. Following recent works, we exploit a large language model to produce the data. The extensive human evaluation shows that dialogue episodes in CONVERSATION CHRONICLES reflect those properties while maintaining coherent and consistent interactions across all the sessions. We also propose a dialogue model, called REBOT, which consists of chronological summarization and dialogue generation modules using only around 630M parameters.

However, although these chatbot models produce human-like fluent responses, they seem to have a limited ability that only understands short-term dialogue context, making them less applicable in real-world scenarios in which long-term conversational situations are often encountered. Specifically, they do not care about the context of past conversations and only generate responses based on an ongoing dialogue (so-called single-session dialogue).

Time interval plays an important role to infuse dynamics in a conversational interaction between speakers. For instance, depending on the time elapsed since the last conversation, their responses about past events would vary. However, previously introduced works have a relatively short range of time intervals, limiting types of transitions from the past sessions.