Overview
Ph.D. candidate at McGill University and Mila, exploring key challenges at the intersection of robotics and AI. My research focuses on reliable policy learning for robotic systems, drawing from world modeling, reinforcement learning, diffusion modeling, vision-language-action models, and multimodal learning to develop intelligent systems that operate safely and reliably to solve complex dexterous manipulation challenges in the physical world.
Education
McGill University & Mila
Ph.D. candidate in Electrical and Computer Engineering, focused on robot learning — GPA: 4/4
Sharif University of Technology
B.Sc. in Electrical Engineering, focused on digital systems — GPA: 3.7/4
Minor in Computer Science, focused on machine learning — GPA: 3.8/4
Minor in Computer Science, focused on machine learning — GPA: 3.8/4
Experience
Researcher — McGill University, Mila – Quebec AI Institute
Tackling the challenge of learning robot policies that are both effective and provably safe in real-world environments. Developed methods that combine reinforcement and imitation learning with fundamental control-theoretic certificates to design policies with formal safety and stability guarantees.
Visiting Fellow — École Polytechnique Fédérale de Lausanne (EPFL)
Addressed the challenge of out-of-sample error accumulation in imitation learning. Developed a framework using neural ordinary differential equations combined with contraction theory to ensure robust imitation and enhanced out-of-sample recovery.
MITACS Research Intern — Sycodal Electronics Inc.
Safe and robust reinforcement learning for industrial manipulation tasks. Designed customized simulation environments in Isaac Sim and applied domain randomization to train robust policies in Isaac Lab. The resulting approach enabled adaptable manipulation in realistic industrial settings.
Researcher — Max Planck Institute for Intelligent Systems
Developed an agent-based probabilistic simulation engine to model epidemic outbreak patterns and assess the effectiveness of various control policies. Learned to control a tri-finger manipulator platform through model-free reinforcement learning to enable object manipulation from vision feedback.
Student Researcher — Artificial Creatures Lab
Addressing the problem of finding diseased plants in vertical farming through ResNet-based vision architectures and acting on the output of a fuzzy control system.
Selected Publications
ECCV 2026
Tactile Modality Fusion for Vision-Language-Action Models
Charlotte Morissette, Amin Abyaneh, Wei-Di Chang, Anas Houssaini, David Meger, Hsiu-Chin Lin, Jonathan Tremblay, Gregory Dudek
ICLR 2026
Contractive Diffusion Policies: Robust Action Diffusion via Contractive Score-Based Sampling
Amin Abyaneh, Charlotte Morissette, Mohamad Danesh, Anas Houssaini, David Meger, Gregory Dudek, Hsiu-Chin Lin
ICLR 2025
Contractive Dynamical Imitation Policies for Efficient Out-of-Sample Recovery
Amin Abyaneh*, Mahrokh Boroujeni*, Hsiu-Chin Lin, Giancarlo Ferrari-Trecate
ICRA 2024
Globally Stable Neural Imitation Policies
Amin Abyaneh, Mariana Sosa Guzmán, Hsiu-Chin Lin
CoRL 2023
Learning Lyapunov-Stable Polynomial Dynamical Systems Through Imitation
Amin Abyaneh, Hsiu-Chin Lin
Full publication list on Google Scholar.
Current Projects
Contact-Aware World Models
Vision-only world models lack the necessary grounding in contact dynamics. We attempt to fix that by capturing contact forces and fusing them with vision-prediction world models.
Drift Q-Learning
Using mean-flow methods instead of denoising diffusion and flow policies in offline continuous control settings, with state-of-the-art results across most environments. NeurIPS 2026 submission.
Technical Skills
Programming
Python, C/C++, Bash; PyTorch, JAX, CUDA/Triton; Hugging Face (Transformers, Diffusers, LeRobot); distributed training (DDP/FSDP), Docker, Slurm, Hydra, Weights & Biases
Machine Learning
Reinforcement, imitation, and offline RL; diffusion and flow matching (mean-flow); world models; vision-language-action models; vision transformers; representation, self-supervised, and multimodal/tactile learning; neural ODEs and SDEs
Robotics & Control
Dexterous and contact-rich manipulation; sim-to-real transfer and domain randomization; stability and safety certificates (Lyapunov, contraction); model-based control, kinematics and dynamics, system identification, motion planning
Robots & Sim
Isaac Sim and Isaac Lab, MuJoCo (MJX), Genesis, PyBullet, ROS 2 · Kinova (Link6, Gen3), Franka Panda, ANYbotics (ANYmal C), Robotiq and Franka grippers, DIGIT tactile sensors
Recent Awards
Fonds de recherche du Québec, Nature et technologies (FRQNT) Doctoral Scholarship
4-year fellowship awarded to distinguished PhD candidates in Quebec.
NCCR Automation Fellowship
Awarded by the Swiss National Centers of Competence in Research to selected international researchers.
MITACS Accelerate Fellowship
Awarded by MITACS Canada to conduct high-caliber research with an industrial partner.
Thomas and Penelope Deirdre Szirtes Fellowship in Engineering (SFE)
Awarded by the Faculty of Engineering on the basis of academic merit to graduate students.
McGill Graduate Excellence Fellowship (GEF)
Awarded for four consecutive years for excellent academic performance.
Max Planck Society Research Scholarship
Awarded for independent research to visiting and international doctorate students.
McGill Engineering Doctoral Award (MEDA)
A competitive scholarship with an acceptance rate below 2 percent for direct admissions.
Fellowship of Iran's National Elites Foundation
Dedicated to the top 0.05 percent of Iran's university entrance exam participants.
Teaching Assistant
McGill University: Applied Robotics (Head TA), Intelligent Robotics (Head TA), Linear Systems, Foundation Models in Robotics (Head TA)
Sharif University: Digital Circuits & Computer Architecture (Head TA), Machine Learning (Head TA)
Leadership and Mentorship
McGill University
Mentoring and advising 4 MSc students and 2 PhD students.
McGill, Columbia, NVIDIA, Amazon
Initiating and leading large-scale collaborations in world models research, including multiple industry/academia projects with NVIDIA and Amazon Research.
Mila
Organizing the World Modeling Workshop at the Reinforcement Learning Conference (RLC) 2026.
Conference Reviewer
Conference on Neural Information Processing Systems (NeurIPS)
2025, 2026
International Conference on Learning Representations (ICLR)
2024, 2025
Conference on Robot Learning (CoRL)
2024
IEEE Robotics and Automation Letters (RA-L)
2024
IEEE International Conference on Robotics and Automation (ICRA)
2023, 2024, 2025
Conference on Robots and Vision (CRV)
2023, 2024
IEEE-RAS International Conference on Humanoid Robots (Humanoids)
2023
Presentations
IEEE International Conference on Robotics and Automation (ICRA) 2024
Yokohama, Japan
Conference on Robot Learning (CoRL) 2023
Georgia, United States
Languages
English
Test of English as a Foreign Language (TOEFL): 111 / 120
French
Basic knowledge (A2)
Voluntary Activity
Computer Science Graduate Society (CSGS), McGill University
Member.
Setak NGO
Member. Countering child labor and helping victims access proper education.
Hackathon 2017, Sharif University
Organizer of a student-run event to accelerate start-ups.
Resana, Student Society of the Electrical Engineering Department, Sharif University
Member.