Collaborative Artificial Intelligence – CSE3210 (Bachelor Computer Science and Engineering, 3rd year, Elective)

Traditionally, Artificial Intelligence (AI) applications are realized as “centralized” solutions, where a single entity, e.g., an organization, collects massive amounts of data from end users to provide a suite of AI-powered services. The ethical concerns such centralized AI solutions pose (e.g., fairness, accountability, transparency, and privacy) are increasingly evident. This course introduces collaborative artificial intelligence, which facilitates interactions among intelligent agents, including humans and AI agents, to realize decentralized AI applications. The topics covered include: Co-active design of decentralized AI systems; Negotiation among autonomous agents; Agent communication and interaction protocols; Agent coordination mechanisms.

In this course, I teach about computational social choice as an approach toward agent coordination and Responsible AI.

Artificial Intelligence Techniques CS4375)

The course focuses on concepts and techniques from probabilistic Artificial Intelligence for building intelligent agents, and decision making and -support systems. The specific topics include Bayesian networks, inference, belief monitoring, (partially observable) Markov decision processes, reinforcement learning and multiagent systems.

In this course, I teach about Reward Hacking and Inverse Reinforcement Learning.