Mohamad H Danesh

prof_pic.png

I’m a PhD student at McGill University, where I’m being guided by Professor Hsiu-Chin Lin. My work revolves around ensuring safe learning for robot motion planning.

Before coming to McGill, I worked as a research engineer at Ivy. Our goal at Ivy was to bring together popular machine learning frameworks like TensorFlow, PyTorch, Jax, and Numpy. To achieve this, we developed tools that could translate complex graphs from different frameworks into actions that specific devices could carry out. Earlier in my career, I spent time as a visiting researcher at the AdaComp lab in the National University of Singapore. There, I collaborated with Professor Panpan Cai and Professor David Hsu on improving planning for self-driving vehicles when faced with uncertain situations. I also hold a master’s degree from Oregon State University. During my time there, I was fortunate enough to be mentored by Professor Alan Fern on explainable and robust RL.

My primary research interest lies in RL. Over time, I’ve explored various aspects of this field, including making RL explanations clearer, ensuring it remains reliable and safe, and effectively planning and learning when outcomes are uncertain. Currently, I’m directing my efforts towards merging RL techniques with robotics.

I’m open to collaborating with others. Feel free to reach out via email if you have any questions regarding an RL project, or if you’re interested in a research collaboration. Also, you may want to take a look at the get in touch page.

Selected Publications

2024

  1. AAAI Oral
    Contextual Pre-Planning on Reward Machine Abstractions for Enhanced Transfer in Deep Reinforcement Learning
    Guy Azran, Mohamad H Danesh, Stefano V Albrecht, and 1 more author
    In 38th Annual AAAI Conference on Artificial Intelligence, 2024
  2. Open RL Benchmark: Comprehensive Tracked Experiments for Reinforcement Learning
    Shengyi Huang, Quentin Gallouedec, Florian Felten, and 8 more authors
    arXiv preprint arXiv:2402.03046, 2024

2022

  1. CoRL Oral
    LEADER: Learning Attention Over Driving Behaviors For Planning Under Uncertainty
    Mohamad H Danesh, Panpan Cai, and David Hsu
    In 6th Annual Conference on Robot Learning, 2022

2021

  1. ICML
    Re-understanding Finite-State Representations of Recurrent Policy Networks
    Mohamad H Danesh, Anurag Koul, Alan Fern, and 1 more author
    In International Conference on Machine Learning, 2021
  2. ICML UDL Oral
    Out-of-Distribution Dynamics Detection: RL-Relevant Benchmarks and Results
    Mohamad H Danesh, and Alan Fern
    In International Conference on Machine Learning, UDL Workshop, 2021