ANIMATAS targets scientific, technical and innovations around three main components :
1) Exploration of fundamental questions relating to the interconnections between robots and virtual characters’ appearance, behaviours and perception by people (WP2).
We designed studies to investigate the relationships between behavioural and appearance qualities of agents and measure their impact on users’ perceptions. We developped computational models for the generation of multimodal behaviours and emotional expressions for social agents in interaction settings.
2) Development of new social learning mechanisms that can deal with different types of human intervention and allow robots and virtual characters to learn in an unconstrained manner (WP3)
We focused on the development of mechanisms for robot/character social learning, i.e. learning in a social context via the interaction with humans. This means that humans act as teachers and robots as learners. Different machine learning techniques are used for robot learning. While robots/characters learn skills for interacting with humans, they also act as educational agents (e.g. tutors, peers) supporting human learning.
3) Development of new approaches for robots and virtual characters’ personalised adaptation to human users in unstructured and dynamically evolving social interactions (WP4)
We developed new computational approaches for adaptive human-agent interactions that are closely coupled with social cognition abilities to achieve mutual co-adaptation in learning scenarios
During the interaction, the agent, be it virtual or physical, could act as a tutor, a peer, a mentor, etc. The interaction may be dyadic or multi-party. Humans (e.g. children) will interact with agents in an education setting, where establishing and maintaining trust for these agents is crucial to ensure collaborative interaction. The different research questions are addressed theoretically and experimentally; computational models are drawn and implemented from these studies and tested in the lab as well as in ecological situations (i.e. schools) whenever possible.
ANIMATAS defined and implemented a structured training programme that leverages cross-disciplinary, cross-national and cross-sectorial strengths.
ANIMATAS made several key contributions at the core of human-machine/robot/agent interaction, and others target agents/robots for learning. The goal was to achieve high academic standards demonstrated by the publication and dissemination of results in top-venues.
