Abstract
There has been a growing interest in empathetic dialogue systems (EDS) that demonstrate emotional and social intelligence within the natural language processing (NLP) community. The ability to recognize users' emotional states, sentiments, and behaviors and respond accordingly is crucial for effective communication. While this comes naturally to humans, embedding such human-like capabilities into conversational systems presents a formidable challenge. Since empathetic dialogue systems encompass a broad range of human characteristics, newcomers often find it challenging to begin. As a result, this tutorial aims to provide a comprehensive overview of empathetic dialogue systems. The tutorial's scope encompasses various aspects, including the historical context, development, available datasets, techniques (including cutting-edge approaches), methods for evaluation, and recent trends and challenges in empathetic dialogue systems. Lastly, it highlights some of the limitations and deficiencies in existing techniques to guide future research in this field.
Outline
The tutorial is organized as follows:
- Concepts of empathy and its need in dialogue systems
- Emotion, sentiment, emotion cause, intent, persona, and politeness in empathetic dialogue systems
- Relevance of external knowledge and multimodal information in empathetic dialogue systems
- Empathetic dialogue systems for social good applications - Persuasion and Therapy
- Conclusion and Future Directions