This dissertation aims to explore the potential of using socially embodied agents to enhance social connections between individuals. The focus is on understanding the social intelligence required for these agents to promote positive human-human interactions and relationships. The research also addresses the responsibility of designing, building, and evaluating computational systems that support a robot’s social intelligence. The increasing use of AI devices in everyday life highlights the need for responsible AI design to ensure human flourishing. The research takes a multidisciplinary approach, developing design frameworks and computational tools for autonomous personalized robot companions. Projects under this research include the study of interpersonal dynamics in dyadic interactions, the design of parent-child-robot interactions in shared reading, the development of a context-generic design framework for multi-person interactions, and the design of affective sensing and behavior personalization models for triadic dyad-robot interactions.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| Increasing availability of AI devices raises concerns about harm to human-human connections | From reliance on AI devices to responsible AI design for human flourishing | More responsible design and evaluation of AI agents for positive human-human interactions | Societal calling for responsible AI |
| Research focus in HRI needs to expand to multi-person interactions | From single-person HRI to multi-person HRI | Development of conceptual frameworks, design principles, and technical tools for MHRI | Increasing availability of social robots in everyday lives |
| Research on personalized MHRI using autonomous robot companions | From single-person HRI to personalized MHRI with two people | Development of design frameworks and computational tools for personalized MHRI | Need for fully autonomous personalized robot companions in social interactions |
| Collection and analysis of parent-child dyad multimodal dataset | Understanding the interpersonal dynamics in dyadic interactions | Importance of accounting for nonverbal behaviors in predicting relationship characteristics | Examining the relationship between nonverbal behaviors and relationship characteristics |
| Design and implementation of parent-child-robot interaction paradigm | Investigating effects of triadic reading on socio-affective connections and reading behaviors | Development of human-centered robot companions for parent-child story time | Understanding the effects of triadic reading and robot behavior strategies |
| Proposal of context-generic design framework for long-term adaptive multi-person interactions | Development of ADAPT-MHRI design framework for MHRI | Integration of robot behavior design and adaptation components, consideration of group-level and individual-level factors | Unifying and extending key concepts in MHRI |
| Design of affective sensing and behavior personalization models for triadic dyad-robot interactions | Personalization of robot behavior in multi-person context | New evaluation methods for analyzing long-term personalization effectiveness | Understanding the impacts of interaction contexts on human-robot dynamics |