Skip to Content

Sriram Subramanian

Sriram Subramaniam headshot

Canada Research Chair in Artificial Intelligence

“I aim to make multi-agent reinforcement learning, or MARL, applicable to real-world problems by improving its sampleefficiency, scaling MARL algorithms to environments with many agents, and investigating its effectiveness in potential practical application domains.”

In the field of artificial intelligence, multi-agent reinforcement learning (MARL) provides algorithms for autonomous agents to learn effective sequential decision-making policies in shared environments. Despite its potential, MARL has yet to be widely used in large-scale real-world decision-making challenges such as fighting wildland fires or autonomous driving. 

Sriram Subramanian of the School of Computer Science will help advance the field of MARL by designing novel algorithms to assist in transferring knowledge and accelerating MARL training, as well as resolving the drawbacks of existing designs to improve algorithm scaling. The potential for these newly designed algorithms to be deployed in complex, real-world environments will be explored.

Learn more about Subramanian’s CRC appointment.

View Subramanian’s full profile.