Interactive systems based on artificial intelligence that aim to support the human in conducting activity, and adapt to an individual and the situation, require new methods for designing, developing and evaluating its behavior. The Collaborative AI Lab provides the environment for multidisciplinary research collaboration on new theories, methods and technology for human-adapted interaction and human-AI collaboration.
Researchers and students from different fields come together in the Collaborative AI Lab to experiment on various topics related to interacting with rational software agents embedded in intelligent systems and environments for work, daily living activities or leisure activities.
Research on human-agent interaction and collaboration includes user modelling, modeling human activity, including its context, environment, and goal models, and methods for detecting intentions. Research also includes methods for collaborating on generating new knowledge, reaching agreements, and deciding what to do, e.g. about a medical condition. Formal dialogue systems as well as other interfaces for human-AI-collaboration are explored both in the Collaborative AI Lab and in people's daily activities.
A participatory action research methodology and co-design are applied, where evaluation of prototypes using formative and situated methods is central. Stakeholders and end users representing future use environments are actively contributing to projects. A major goal is to facilitate autonomy and optimize skills in individuals. Consequently, potential end users and stakeholders play a key role in the projects and contribute to the multidisciplinary research environment. They enable and realize the user-driven development methods and the technology that we develop by developing knowledge-based support applications following their envisioned future practice/activity.
Location and Resources
The Collaborative AI Lab can be found in room C420 at the top floor in the MIT Building and is maintained by the research group on Interactive and Intelligent Systems. The lab contains a home environment with interaction technology such as sensors, NAO robots, Hololense and other equipment, and a number of prototype applications that have been developed as part of research for different purposes related to activity and complex decision-support in every day/work activities.
The research and innovation platform ACKTUS is used as research infrastructure for co-design and knowledge engineering, to allow experts in various domains to hands-on develop their knowledge-based systems.