Researcher Aims to Equip AI Assistants with Empathy at University of Stuttgart

编辑者: Elena Sun HealthEnergy

Andreas Bulling is researching intelligent assistance systems at the University of Stuttgart, designed to provide optimal support in fields such as medical diagnostics, caregiving, and everyday tasks. In the future, these systems are expected to understand and empathize with users. Bulling has been named a Henriette Herz Scout by the Alexander von Humboldt Foundation for his expertise in human-machine interaction and commitment to nurturing early career researchers.

Intelligent assistance systems will increasingly collaborate with humans in various settings, including healthcare and daily life. Currently, they can recognize body language, facial expressions, and eye movements, detecting emotions like sadness or happiness. However, these systems still lack the ability to fully understand another person's perspective or respond empathetically.

“To achieve this, we must teach these systems to recognize our needs, goals, and intentions,” says Prof. Andreas Bulling, an expert in human-machine interaction and cognitive systems at the University of Stuttgart. This requires significantly improved sensory and cognitive capabilities. “The vision is to create intelligent assistants or avatars that we can interact with as if they were human,” he adds.

Bulling has nominated three Humboldt Research Fellows for a two-year fellowship in his Collaborative Artificial Intelligence (CAI) department at the Institute for Visualization and Interactive Systems (VIS). He emphasizes that this unique opportunity aims to attract young talent to the University of Stuttgart.

Since receiving an ERC Starting Grant in 2018, Bulling has supported 14 doctoral students and postdocs, seven of whom are now professors themselves. Currently, he supervises 16 additional doctoral students in his research group. Bulling's appointment as a Henriette Herz Scout marks him as the second Stuttgart scout, following Professor Johannes Kästner.

He is currently conducting extensive research to enable intelligent agents to develop a theory of mind, which refers to the cognitive ability to empathize with others. This involves understanding mental states as potential causes of observed behavior, allowing for the prediction and explanation of social actions. While most people naturally use this skill in interactions, teaching computers to do the same is complex.

If Bulling and his team succeed, it could lead to the development of assistance systems capable of proactive actions. For instance, such an assistant might recognize when a patient needs help with a daily task and offer support accordingly. This would allow assistance systems to respond to the emotional states and needs of their human counterparts.

In Stuttgart, Humboldt Fellows benefit from a comprehensive computer science department that collaborates with two Clusters of Excellence—SimTech and IntCDC—and the ELLIS Unit, which Bulling co-leads. “We are among the few interdisciplinary research groups in Germany and Europe specializing in human-machine interaction and cognitive modeling,” he notes.

In addition to research, the fellows will gain experience in teaching and supervising doctoral students, as well as acquiring external funding. “The development of young scientists is very important to me, which is why I provide comprehensive and engaged support,” says Bulling.

The Henriette Herz Scouting Program is aimed at researchers at all career stages and from various disciplines across Germany. Candidates must hold a (junior) professorship or a comparable leadership position and have an established international collaboration network. Scouts can nominate up to three international scientific talents for a two-year research fellowship, with up to 100 scholarships awarded annually through a direct selection process.

For further information, Prof. Andreas Bulling can be reached at the University of Stuttgart, Institute for Visualization and Interactive Systems, Tel.: +49 711 685 60048, email: andreas.bulling@vis.uni-stuttgart.de.

你发现了错误或不准确的地方吗?

我们会尽快考虑您的意见。