PersonalityChat: Conversation Distillation for Personalized Dialog Modeling with Facts and Traits
This work addresses the problem of costly and difficult crowd-sourcing for personalized dialog datasets, offering a method to enhance conversational AI personalization, though it is incremental by building on existing datasets and methods.
The authors tackled the challenge of creating personalized conversational datasets by using LLMs to generate PersonalityChat, a synthetic dataset based on PersonaChat but with added personality traits, and found that models fine-tuned on it enabled trait-based personalization and produced more fluent and coherent dialog agents in small models compared to PersonaChat.
The new wave of Large Language Models (LLM) has offered an efficient tool to curate sizeable conversational datasets. So far studies have mainly focused on task-oriented or generic open-domain dialogs, and have not fully explored the ability of LLMs in following complicated prompts. In this work, we focus on personalization, and employ LLMs to curate a dataset which is difficult and costly to crowd-source: PersonalityChat is a synthetic conversational dataset based upon the popular PersonaChat dataset, but conditioned on both personas and (Big-5) personality traits. Evaluating models fine-tuned on this dataset, we show that the personality trait labels can be used for trait-based personalization of generative dialogue models. We also perform a head-to-head comparison between PersonalityChat and PersonaChat, and show that training on the distilled dataset results in more fluent and coherent dialog agents in the small-model regime.