About Me
Hello! My name is Amit Sinha and I’m currently in the final year of my PhD at McGill University under the supervision of Prof. Aditya Mahajan. I’ve also previously worked with Matthieu Geist, Jayakumar Subramanium and Sarath Chandar.
Research Interests
My current research interests are centered around reinforcement learning for partially observable systems and multi-agent team-based systems.
In other words, I consider how vast amounts of complex information used by an automated decision making agent can be reduced to smaller, simpler and more efficient blocks of information; and subsequently how we can use these efficient blocks of information to learn automated decision making strategies that are useful.
Recent News
Jun 2025: Excited to be doing a PhD research internship at Adobe during the summer!
Dec 2024: Attending NeurIPS 2024.
Dec 2024: Attending CDC 2024.
Sep 2024: Our paper: Periodic agent-state based Q-learning (PASQL) for POMDPs was accepted in NeurIPS 2024 in Vancouver!
Jul 2024: Our tutorial paper: Agent-state based policies in POMDPs: Beyond belief-state MDPs was accepted in CDC 2024 in Milan, Italy!