About me
Machine Intelligence and Robot Learning (MIRL) group:
Building safe, data-efficient AI systems for real-world deployment.
I am an Associate Professor of Computer Science at the University of Ottawa (EECS) and a Faculty Affiliate with the Vector Institute. My research focuses on deep learning and reinforcement learning for real-world systems, with an emphasis on safety, efficiency, and learning under constraints.
I study how intelligent agents can learn from limited, costly, or uncertain data, and safely adapt to evolving environments. My work spans goal-conditioned learning, safe decision-making, and data-efficient machine learning, with applications in robotics, healthcare, the environment and scientific discovery.
I lead the Machine Intelligence and Robot Learning (MIRL) group and am a member of the Instrumentation and Autonomous Robotics Agents IARA lab.
Research Areas
- Reinforcement Learning: safe RL, model-based RL, sample efficiency, sim-to-real
- Deep Learning: imbalanced/limited data, active learning, interpreability
- Applications: robotics, healthcare, industrial automation, AI for science
Latest News
- May 2026 — Two papers accepted at RLC 2026
- April 2026 — Awarded NSERC Discovery Grant. See the research summary here and proposal here.
- April 2026 — MSc thesis defense: Alireza Seyed Azimi
- April 2026 — Honours research presented at Canadian AI 2026
Prospective Students
I am always looking for motivated students interested in reinforcement learning, deep learning, and robotics. If you are interested please applied to the appropriate graduate program and completed the linked form here.
For MSc/PhD applicants: please list me as your preferred supervisor. I review applications after admission decisions (spring intake for September start).
Selected Publications
View full publication list (Google Scholar)
- General and Efficient Visual Goal-Conditioned Reinforcement Learning using Object-Agnostic Masks, RLC, 2026
- Deep Policy Gradient Methods Without Batch Updates, Target Networks, or Replay Buffers, NeurIPS, 2024
- ChemGymRL: A customizable interactive framework for reinforcement learning for digital chemistry, Digital Discovery, 2024
- Understanding CNN fragility when learning with imbalanced data, Machine Learning, 2024
- Learning When to Observe: A Frugal Reinforcement Learning Framework for a High-Cost World, ECML-PKDD, 2023
- The class imbalance problem in deep learning, Machine Learning, 2022