EmpDG: Multi-resolution Interactive Empathetic Dialogue Generation

Abstract

A humanized dialogue system is expected to generate empathetic replies, which should be sensitive to the users' expressed emotion. The task of empathetic dialogue generation is proposed to address this problem. The essential challenges lie in accurately capturing the nuances of human emotion and considering the potential of user feedback, which are overlooked by the majority of existing work. In response to this problem, we propose a multi-resolution adversarial model -EmpDG, to generate more empathetic responses. EmpDG exploits both the coarse-grained dialogue-level and fine-grained token-level emotions, the latter of which helps to better capture the nuances of user emotion. In addition, we introduce an interactive adversarial learning framework which exploits the user feedback, to identify whether the generated responses evoke emotion perceptivity in dialogues. Experimental results show that the proposed approach significantly outperforms the state-of-the-art baselines in both content quality and emotion perceptivity.

Modelling Context Emotions using Multi-task Learning for Emotion Controlled Dialog Generation

“…polite, rude or neutral). Li et al (2019) proposed an empathetic dialogue system (EmpGAN) based on adversarial learning comprising of a multi-resolution empathetic generator along with two interactive discriminators. Song et al (2019) presented an attention framework based on emotion-lexicons.…”

Anna: A Dapper Open-Domain Dialogue Agent Based on a Joint Attention Network

“…Some studies have improved the anthropomorphic characteristics of generated responses. For example, the backgrounds of persona (Zhang et al 2018;Mazaré et al 2018;Madotto et al 2019;Cui et al 2018) and emotional information (Zhou et al 2020;Huang et al 2018;Rashkin et al 2018;Li et al 2019) have been incorporated into encoder-decoder architectures. These studies aimed to address the problem of generating generic responses, which is also one of the motivations for our study.…”

Empathetic Response Generation through Graph-based Multi-hop Reasoning on Emotional Causality

“…MIME [5] assumes that empathetic responses often mimic the emotions of the speakers to some degree, and its consideration of polaritybased emotion clusters and emotional mimicry shows to improve empathy and contextual relevance of responses. Em-pDG [6] exploits both dialogue-level and token-level emotion to capture the nuances of the speaker's emotion. Moreover, it also involves the speaker's feedback to achieve better expression of empathy with an interactive adversarial learning framework.…”

“…Empathetic responding, which aims to understand the emotional experience of the user and then reply to acknowledge it appropriately, has benefited a wide range of downstream applications, such as medical dialogue systems [1], counseling conversation [2] and social chatbots [3]. Recent years have witnessed the emergence of empathetic dialogue systems, and some methods [4,5,6] have been proposed for the task of empathetic response generation and achieved promising results. Specifically, to better understand the emotional experience of the user, MoEL and MIME [4,5] utilize the emotional labels, and EmpDG [6] uses both coarsegrained labels annotated by human and fine-grained emotional words identified by the sentiment lexicons.…”

“…Recent years have witnessed the emergence of empathetic dialogue systems, and some methods [4,5,6] have been proposed for the task of empathetic response generation and achieved promising results. Specifically, to better understand the emotional experience of the user, MoEL and MIME [4,5] utilize the emotional labels, and EmpDG [6] uses both coarsegrained labels annotated by human and fine-grained emotional words identified by the sentiment lexicons. However, these methods only focus on what the emotion is when encoding the conversation history, which is insufficient to fully understand the emotional experience of the user.…”