DuLeMon

该数据集名为DuLeMon，是最新开放的具有长期个性记忆的对话数据集。在DuLeMon中，回应是基于历史会话中出现的个性信息，这有助于提高对话的参与度。DuLeMon有两个版本：Self和Both。在Self版本中，个性信息仅来自单方面；而在Both版本中，则考虑了用户和聊天机器人双方独立的个性信息。相关论文采用了DuLeMon的Both版本进行个性化对话生成任务的研究。The dataset named DuLeMon is a newly open-sourced conversational dataset equipped with long-term personality memory. In DuLeMon, responses are generated based on the personality information presented in historical conversations, which effectively boosts conversational engagement. DuLeMon offers two variants: Self and Both. For the Self variant, personality information is exclusively sourced from a single party; whereas for the Both variant, independent personality information from both the user and the chatbot is taken into consideration. A relevant research paper adopted the Both variant of DuLeMon to conduct studies on personalized dialogue generation tasks.

搜集汇总

背景与挑战

背景概述

DuLeMon是一个最新开放的对话数据集，专注于长期个性记忆，通过基于历史会话中的个性信息生成回放，以提升对话参与度。它包含两个版本：Self版本仅涉及单方面个性信息，而Both版本则整合了用户和聊天机器人双方的独立个性信息，相关研究采用Both版进行个性化对话生成任务。