I want to teach AI agents how to selectively attribute success or failure to the decisions taken in the past. In other words, I am interested in solving the temporal credit assignment problem in reinforcement learning.
In the future, I believe AI can solve many problems across several domains and make our lives easy. I also like to think that we would be synergistically living with AI agents in the future.
I completed my MSc in Computer Science at McGill University before starting a Ph.D. I worked on solving temporal credit assignment through traces as a part of my thesis.