Building a Personalized Dialogue System with Prompt-Tuning
This addresses the need for consistent and personalized dialogue systems, but it is incremental as it applies an existing method (prompt-tuning) to a specific domain.
The authors tackled the problem of inconsistent responses in dialogue systems by building a system that responds based on a given persona, using prompt-tuning on pre-trained large language models. The results show it achieves more natural and personalized responses with less computational resources than fine-tuning, as demonstrated in automatic and manual evaluations in English and Japanese.
Dialogue systems without consistent responses are not fascinating. In this study, we build a dialogue system that can respond based on a given character setting (persona) to bring consistency. Considering the trend of the rapidly increasing scale of language models, we propose an approach that uses prompt-tuning, which has low learning costs, on pre-trained large-scale language models. The results of automatic and manual evaluations in English and Japanese show that it is possible to build a dialogue system with more natural and personalized responses using less computational resources than fine-tuning.