Dialogue Agents with Literary Character Personality Traits

Junzhe Zhao, Huizhi Liang, Nicolay Rusnachenko

Abstract: To enhance the engagement of chatbots, we need to imbue them with unique personalities and speech patterns. However, creating a high-quality conversational training dataset for this task can be time-consuming and labor-intensive. To address this, we propose an automated system that uses character dialogues from literary works. We used the Chinese classic, (Dream of the Red Chamber), as our data source. Our system efficiently extracts dialogues and personality traits from the book, creates a personality map for each character, and generates responses that reflect these traits. It allows us to utilize the distinct linguistic styles that authors assign to characters with different personalities in classic books, making chatbots more human-like. Our evaluations demonstrate the efficacy of our system, with chatbots possessing personality traits proving more engaging than those lacking them. For transparency, we have made all our datasets and models available ¹ ¹ Project Repository: https://github.com/SuperEDG/Hongloumeng_Project.