First call for proposals results in promising new Human-centered AI research collaborations

Cognition is generally accepted as a crucial factor for independence in various activities in daily life and participation in society across the lifespan. After brain damage at young age, cognitive development (e.g. attention, executive functions, memory, language, social cognition) can be seriously hampered, sometimes even up to the level of being completely dependent in daily activities and no active participation.

These problems in cognitive development and their consequences have a profoundly negative impact on quality of life. Better understanding of learning and learning potential, and the maximum capacity of learning, given the type of brain damage, will lead to better models for both predicting cognitive development and explaining cognitive complaints.

Building a cognitive model for the developing brain that could be adjusted to diagnosis specific characteristics and cognitive domains would be a crucial next step for all the fields of research. For such an endeavour, a multidisciplinary approach is required, bringing together experts from cognitive modelling, cognitive neuroscience and medical doctors.

By organising a seminar, the applicants would not only bring together all researchers from different collaborations, but would also introduce everyone involved to the potential of computational modelling in healthcare. As such, this project would add new collaborators from the field of HAI to the existing collaborations. One important aim of this seminar is to connect the focus area with Prinses Maxima Center and the Wilhelmina Children's Hospital (WKZ). There is a great promise in involving these centres as stakeholders and active participators in the newly founded focus area in an early stage.

The project will combine philosophy of logic and computing sciences to investigate the normative constraints logic may put on reasoning of both human agents and AI systems. The team aims to contribute to the resolution of notorious problems within the research field like, for instance, the challenges related to logical omniscience and the lack of guidance logic offers concerning the question of which propositions to accept.

The team will organize an international workshop, which will serve as a kick-off event to establish connections between the involved disciplines and explore potential questions for future interdisciplinary research. The organisers come from two different faculties and will bring their complementary perspectives on the nature of reasoning and its implementation in AI systems.

Their contacts in the respective research communities will produce an inclusive event with a broad interdisciplinary appeal. The proposed workshop will be organised under the auspices of HAI and will bring to Utrecht researchers from other institutions in the Netherlands and from universities elsewhere in the world to work on issues closely connected to Human-centered AI.

The modeling of emotional intelligence has become a vivid theme in recent years among the areas of philosophy, psychology, computer science, and artificial intelligence. Emotions serve essential functions for human to make judgments in reasoning and decision making. For instance, to prevent experiencing regret, the choice of investing a risky high-tech fund might be dissuaded by this expected emotion.

Emotions coordinate human social interaction because of the informational value and the communicative roles involved in them, which is still scarcely explored in formal work. To model emotional intelligence in social interaction, on the one hand, we need to encode emotions as cognitive component to be integrated in the reasoning process. Emotional regulation characterizes which actions deserve to take as social responses to the current situation. Regarding the social effect of emotion, the agents should consider the potential social consequences which motivate their affective behaviors.

The goal of this collaborative research is to model the reasoning patterns of emotions in social interaction. In particular, our research pays attention to the strategic behaviors of human-centered emotional intelligence. As the studies in behavioral science and psychology show, some social situations offer more cooperative than competitive incentives and thereby drive participants toward cooperation; otherwise, toward competition. We plan to develop a formal model of human-centered reactive system that counts for emotional intelligence and reasoning in social interaction.

The applicants will invite postdoctoral researcher Dr. Huimin Dong from Zhejiang University, China to join the research team and contribute to the logical development of human-centered AI. Based on this collaboration, the team intends to organize a seminar and submit joint work for a conference.