Artificial Intelligence (AI) methods have advanced tremendously in the past several years and it is not uncommon to see them match or surpass human performance in tasks with well-defined objectives. However, most relevant real-world problems, specifically those that involve humans as users or active agents, are difficult to fully specify a priori.

One principled way to address this issue is for AI systems to collaborate with humans by actively reasoning about and anticipating human needs and abilities. Human-AI collaboration can be highly effective if the AI system considers human needs and adapts to them on the fly — however, this would require realistic computational models of the human user. Computational models of human behavior have been heavily investigated in cognitive science as well as in other fields including human-computer interaction (HCI), human-robot interaction (HRI), etc. However, due to their different goals and experimental settings, many relevant areas of research have operated more or less independently, with sub-optimal communication between them.

In this bridge program, we aim to bring together members of the communities relevant to human-AI collaboration and user modeling to create immediate and long-term interdisciplinary projects. This bridge will expose AI researchers to methods and advances in methods for user modeling, opening the door to solving problems with human interactions. In turn, modelers outside traditional AI will be exposed to new applications, prospects, and techniques that are relevant for user models, new learning and inference techniques, as well as new potential for advances in state-of-the-art user modeling through techniques from AI.

For Attendees!