Matin Moezzi | M.A.Sc. U of Toronto

About Me

I’m a machine learning researcher exploring Reinforcement Learning (RL), focusing on sample efficiency and generalizability to improve real-world applications. Moving forward, I’m driven to advance Physical AI, particularly in robotics, to bring AI into tangible systems.

Research Interests

Model-based RL, Meta-RL, Offline RL
Diffusion Models, Score-based Generative Models
Imitation Learning, Sim2Real

News

[Sept 2024] Excited to join ArenaX Labs as a ML Researcher, leveraging RL to advance toward superintelligence!”
[Jan 2024] I successfully defended my MASc thesis at the University of Toronto: “Towards Sample-Efficient Reinforcement Learning Methods for Robotic Manipulation Tasks”
[Aug. 2023] Our paper titled “A Comparison of Classical and Deep Reinforcement Learning Methods for HVAC Control” is accepted to IEEE SWC 2023.
[Jan. 2023] I joined Vector Institute as an Applied Machine Learning Intern.
[May 2022] I was admitted to the NSERC Alliance-Mitacs Accelerate Internship program.

Publication

Towards Sample-Efficient Reinforcement Learning Methods for Robotic Manipulation Tasks

Matin Moezzi

University of Toronto

PDF
A Comparison of Classical and Deep Reinforcement Learning Methods for HVAC Control

Marshall Wang, John Willes, Thomas Jiralerspong, Matin Moezzi

IEEE Smart World Congress 2023

PDF Code
An uncertainty-aware pseudo-label selection framework using regularized conformal prediction

Matin Moezzi

arXiv

PDF

Research Experiences

Applied Machine Learning Intern

Vector Institute, Jan - Apr 2023

PDF Code

DynamicsDiffusion: Model-based RL with Score-based Diffusion Models

Graduate RA, University of Toronto, Jan 2022 - Present

End-to-End Sim-to-Real Transfer Learning

Graduate RA, University of Toronto, Jan 2022 - Present

Code

Mitacs Accelerate Intern

University of Toronto, May 2022 - Aug 2022

Code

Teaching Experiences

Graduate Teaching Assistant, University of Toronto

CSC2516: Neural Networks and Deep Learning [Winter by Prof. Jimmy Ba, Fall 2023]
MIE567: Dynamic and Distributed Decision Making [Winter 2023]
APS1080: An Introduction to Reinforcement Learning [Fall 2022, Winter, Summer 2023]
MIE1615: Markov Decision Processes [Winter 2023]
Probability [Fall 2022, 2023]
CSC369: Operating Systems [Fall 2022]

Selected Courses and Certificates

University of Toronto

Imitation Learning for Robotics - A
Probabilistic Learning & Reasoning - A+
Markov Decision Processes - A
Cloud-Based Data Analytics - A

Coursera

Reinforcement Learning Specialization, University of Alberta - See the Certificate
Deep Learning Specialization, Andrew Ng, deeplearning.ai - See the Certificate
Natural Language Processing Specialization, deeplearning.ai - See the Certificate
TensorFlow Developer Specialization, deeplearning.ai - See the Certificate