RECAP: Retrieval-Enhanced Context-Aware Prefix Encoder for Personalized Dialogue Response Generation

Understand

Endowing chatbots with a consistent persona is essential to an engaging conversation, yet it remains an unresolved challenge.

• In this work, we propose a new retrieval-enhanced approach for personalized response generation.
• Specifically, we design a hierarchical transformer retriever trained on dialogue domain data to perform personalized retrieval and a context-aware prefix encoder that fuses the retrieved information to the decoder more effectively.
• Extensive experiments on a real-world dataset demonstrate the effectiveness of our model at generating more fluent and personalized responses.