Image
Image

MARKUS EGER

Dr. Markus Eger is currently an assistant professor in the Computer Science Department at Cal Poly Pomona, and previously was a visiting professor at the University of Costa Rica in the School of Computer Science and Informatics. His research group is called SLOTHLab, which is short for Symbolic Logic Operations and Transactions with Humans. Together with his students, Markus does research on interactions between humans and AI agent, using video games as an application domain. He got his PhD from NC State University in 2018, working on a dissertation titled "Intentional Agents for Doxastic Games", which laid the foundations for his ongoing work on how to build AI agents that communicate effectively with human collaborators. This work combines applications of AI techniques with work on human communication and psychology, and has been published at leading game research conferences such as AIIDE, CIG, or FDG. As an active member of this community, Markus has also served on the program committee for AIIDE, FDG, and several of their satellite workshops.

Topic: AI meets Humans

AI research has made incredible advances in recent years, with applications as varied as image recognition, text classification, mastery of video games, or even creative processes such as text, image or music generation. In several application areas, including Chess or Jeopardy, the performance of AI agents even surpasses that of human experts. However, when the goal for the AI agent is to cooperate, rather than compete, with a human, there are still many challenges that need to be addressed. Human behavior and communication is rich, and based on many implicit and subtle rules that are challenging to reason about.

In this talk I will discuss several challenges we face when humans meet AI, what we can already do, and what has yet to be addressed. I will be placing a particular focus on how games can serve as ideal testing environments to develop new methods that can subsequently be applied to other domains. In addition to traditional, explicit communication, I will also present developments in the interpretation of subconscious communication modes, such as the timing of actions, gaze, and facial expressions.