K-ESConv: Knowledge Injection for Emotional Support Dialogue Systems via Prompt Learning

TL;DR

K-ESConv is a prompt learning method that injects professional forum knowledge into emotional support dialogue systems, significantly improving response quality, diversity, and user comfort in psychological counseling scenarios.

Contribution

This paper introduces K-ESConv, a novel prompt learning approach for knowledge injection from online forums into emotional support dialogue systems, enhancing response relevance and diversity.

Findings

Outperforms existing baselines in automatic and human evaluations

Improves response correlation and diversity

Provides more comforting and helpful suggestions

Abstract

Automatic psychological counseling requires mass of professional knowledge that can be found in online counseling forums. Motivated by this, we propose K-ESConv, a novel prompt learning based knowledge injection method for emotional support dialogue system, transferring forum knowledge to response generation. We evaluate our model on an emotional support dataset ESConv, where the model retrieves and incorporates knowledge from external professional emotional Q\&A forum. Experiment results show that the proposed method outperforms existing baselines on both automatic evaluation and human evaluation, which shows that our approach significantly improves the correlation and diversity of responses and provides more comfort and better suggestion for the seeker.