Personalized And Emotional Conversation

Personalized and Emotional Conversation (PEC) refers to the task of generating human-like responses ($Y$) given the personalized information of two speakers ($P_{R1}$ and $P_{R2}$), multi-turn conversation context ($C$), the target emotion for the response ($E_K$), the dialogue act ($D_K$), and the responder's personalized information ($P_{K}$). This task aims to generate the most appropriate response by maximizing the conditional probability $P(Y'|C, E_K, D_K, P_K)$, thereby enhancing the emotional expression and personalization capabilities of natural language processing systems in human-computer interaction and improving user experience.