Dr. Hossein Hassani
“I want to make safe and effective autonomous systems a reality for applications and users.”
A subject matter expert in embodied artificial intelligence (AI), Dr. Hossein Hassani investigates how to responsibly and sustainably integrate AI with physical systems. While many high-performance AI models exist today, they rely on energy-intensive hardware – creating barriers for users and applications. Inspired to make cutting-edge technologies more accessible, Hassani develops small yet powerful AI models for broader adoption. Throughout his research career, Hassani has specialized in deep reinforcement learning, computer vision, and large language models. Before joining Toronto Metropolitan University as an assistant professor, he held research positions at Western University and the University of Windsor. At TMU, he leads an automotive mechatronics research group, applying algorithmic systems to autonomous vehicles and related technologies. His latest research is focused on developing vision-language-action (VLA) models to enable end-to-end autonomous driving, while prioritizing safety and reliability. |
Highlights
TinyDrive: a novel vision-language (VLM) model for autonomous vehicles
Hassani developed TinyDrive (external link) , a VLM that can achieve real-time visual question-answering on an autonomous vehicle. He’s developing the next generation of the model to push the boundaries of efficiency and generalization, and explore evolving the VLM into a vision-language-action model.
Additional Info
H. Hassani, S. Nikan, and A. Shami, “TinyDrive: Multiscale visual question answering with selective token routing for autonomous driving,” arXiv preprint arXiv:2505.15564, 2025.
H. Hassani, S. Nikan, and A. Shami, “Attention prioritized experience replay with application to self-driving cars,” IEEE Systems, Man, and Cybernetics Magazine, In press.
H. Hassani, S. Nikan, and A. Shami, “Improved exploration-exploitation trade-off through adaptive prioritized experience replay,” Neurocomputing, vol. 614, p. 128836, 2025.
L. Chen, H. Hassani, and S. Nikan, “TS-VLM: Text-guided SoftSort pooling for vision-language models in multi-view driving reasoning,” arXiv preprint arXiv:2505.12670, 2025.
H. Hassani, S. Nikan, and A. Shami, “Traffic navigation via reinforcement learning with episodic-guided prioritized experience replay,” Engineering Applications of Artificial Intelligence, vol. 137, p. 109147, 2024.
|
|
Education
Year |
University |
Degree |
2023 |
University of Windsor |
PhD |
2016 |
Shiraz University of Technology |
MSc |
2013 |
Shiraz University of Technology |
BSc |
Courses
Course Code |
Course Title |
MEC617 |
Manufacturing System Control |