I develop and analyze multi-agent algorithms for solving optimization problems in artificial intelligence (AI), machine learning, and autonomous systems.
Discrete-Time Algorithms
I develop discrete-time optimization algorithms that tolerate asynchrony between agents. Rather than impose assumptions on the environment, I consider certain types of problems that allow for delays between updates and communications among agents. These problems are found in a wide range of settings including machine learning, AI, networks, communications, and robotics.
Hybrid Algorithms
Some applications inherently require a hybrid systems approach which combine continuous-time dynamics and discrete-time events. I develop multi-agent hybrid algorithms that perform gradient descent in continuous-time to solve optimization problems and model communication events between agents in discrete-time.