About

I am currently doing my PhD in Computer Science at Mila (University of Montreal) supervised by Prof. Guillaume Rabusseau. Previously I completed my MASc at the Vision and Image Processing lab at University of Waterloo under the supervision of Prof. David Clausi and Prof. John Zelek. I completed my undergraduate degree at the University of Toronto where I specialized in Mechatronics Engineering.

My research interests include Model Merging, Unlearning and Safety/Alignment. Previously, I worked in computer vision on video action recognition and as well as model compression of large language models.

Selected Publications

Model Merging via Data-Free Covariance Estimation

M. Gamal Abdel Hameed, D. Tam, P. J. Tikeng Notsawo, and 2 more authorsConference on Language Modeling (COLM), 2026

Merging Adapted Models via Data-Free Covariance Estimation Oral

M. Gamal Abdel Hameed, D. Tam, P. J. Tikeng Notsawo, and 2 more authorsICML Workshop on Continual Adaptation at Scale, 2026

A Tensor Decomposition Perspective on Second-order RNNs Spotlight

Maude Lizaire, Michael Rizvi-Martel, M. Gamal Abdel Hameed, and G. RabusseauInternational Conference on Machine Learning (ICML), 2024

ROSA: Random Subspace Adaptation for Efficient Fine-Tuning

M. Gamal Abdel Hameed, Aristides Milios, Siva Reddy, and G. RabusseauICML Workshop on Efficient Systems for Foundation Models, 2023

Generalized Kronecker-based Adapters for Parameter-efficient Fine-tuning of Vision Transformers

A. Edalati, M. Gamal Abdel Hameed, and A. MoslehConference on Computer and Robot Vision (CRV), 2023

Convolutional Neural Network Compression through Generalized Kronecker Product Decomposition

M. Gamal Abdel Hameed, M. S. Tahaei, A. Mosleh, and V. Partovi NiaAAAI, 2022

SeKron: A Decomposition Method Supporting Many Factorization Structures

M. Gamal Abdel Hameed, M. S. Tahaei, A. Mosleh, and V. Partovi NiaarXiv preprint arXiv:2210.06299, 2022

Special Projects

Dekki 50k+ users

A statistical spaced-repetition app for medical students that models individual retention curves to optimize review schedules, built with Luke Kyne.

Papers.app

A semantic search tool for machine-learning papers with conference filtering, built with Jeremy Pinto.

Visualizing Singular Value Decomposition 100k+ views

An animated explanation of the geometric intuition behind SVD, based on key concepts from Gilbert Strang's MIT lecture.

Contact

Email me at marawan [dot] gamal [at] mila [dot] quebec for research collaborations.