Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces PRCCF, a framework for Emotional Support Conversation (ESC) that uses persona-guided retrieval to improve response generation and causality-aware cognitive filtering to prioritize relevant external knowledge. The persona-guided retrieval jointly models semantic compatibility and persona alignment, while the cognitive filtering module enhances contextual understanding for emotional reasoning. Experiments on the ESConv dataset show that PRCCF outperforms existing state-of-the-art baselines on both automatic metrics and human evaluations, suggesting improved empathetic response generation.
Better emotional support chatbots are here: PRCCF leverages persona-aware retrieval and causal filtering to generate more empathetic and contextually relevant responses.
Emotional Support Conversation (ESC) aims to alleviate individual emotional distress by generating empathetic responses. However, existing methods face challenges in effectively supporting deep contextual understanding. To address this issue, we propose PRCCF, a Persona-guided Retrieval and Causality-aware Cognitive Filtering framework. Specifically, the framework incorporates a persona-guided retrieval mechanism that jointly models semantic compatibility and persona alignment to enhance response generation. Furthermore, it employs a causality-aware cognitive filtering module to prioritize causally relevant external knowledge, thereby improving contextual cognitive understanding for emotional reasoning. Extensive experiments on the ESConv dataset demonstrate that PRCCF outperforms state-of-the-art baselines on both automatic metrics and human evaluations. Our code is publicly available at: https://github.com/YancyLyx/PRCCF.