Summary
- Research area: Continual Learning, Federated Learning, Self-supervised Representation Learning, Domain Generalization, Long-tailed Recognition, Biomedical Image Processing, and Neuromorphic Engineering.
- Publications: 13 papers in top venues such as ICASSP, ICIP, and MTAP. More than 370 citations with an h-index of 9.
- Peer review: Regularly reviewing for top venues such as NeurIPS, AAAI, AAAI Workshops, ICML Workshops, IEEE TAI, etc.
- Active Collaboration: with GeoTab Inc., University of Saskatchewan, and the City of Kingston.
- Conference and workshop presentation: Presenting our works at ICASSP, ICIP, and NeurIPS M3L workshop.
- Innovative Contributions: Proposing the first multi-scale capsule GAN, the first continual pedestrian detection method, reformulating and addressing long-tailed recognition as continual learning problem for the first time, and the first federated unsupervised domain generalization solution.
Professional Experience
Graduate Research Assistant
Ingenuity Labs Research Institute, Queen's University
2021 - Present- Collaborated with a team of five researchers to develop innovative methods for crowd-counting, proposing a method that addresses both problems of localization and counting.
- Independently researched and implemented continual learning techniques for pedestrian detection, innovatively adapting a popular continual learning method to pedestrian detection networks, outperforming all existing methods in handling catastrophic forgetting.
- Spearheaded individual research on Continual Learning for Long-Tailed Recognition, proposing a novel framework that unifies continual learning and long-tailed recognition, theoretically analyzing the training under LTR setup and proving the efficacy of our method.
- Engaged in a team-based project focused on unsupervised federated domain generalization, introducing the novel problem of Federated Unsupervised Domain Generalization to the community, proposing a novel solution employing gradient alignment, outperforming all previous methods on all benchmarks.
- Results in 2 publications in ICASSP 2022 and ICIP 2023, and a presented paper in NeurIPS 2023 M3L workshop.
Graduate Research Assistant
Department of Electrical and Computer Engineering, University of Saskatchewan
2019 - 2021- Pioneered the introduction of a multi-scale gradient capsule GAN, advancing super-resolution techniques in both facial and biomedical domains, outperforming SOTA in both fields in all similarity metrics.
- Joined a multidisciplinary team to delve into biomedical image processing, innovatively approaching COVID-19 detection from chest X-ray, introducing a novel public dataset, proposing an end-to-end framework for dental caries detection.
- Results in 4 Q1 journal papers and 2 IEEE conference publications.
Research Intern
Machine Vision Lab, Iran University of Science and Technology
2016 - 2018- Expanded research horizons into neuromorphic systems, designing a novel hybrid CMOS/Memristor circuit-level implementation of Hopfield Neural Network, employing memristor crossbar array for implementing bidirectional associative memory using a low-power low-voltage circuit.
- Collaborated in a dynamic team environment to forecast anomalies in communication networks.
- Results in 2 journal papers and 1 IEEE conference publication.
Education
Queen's University
PhD in Electrical and Computer Engineering
May 2021 - Present | GPA: 4.3/4.3Advisors: Dr. Michael Greenspan and Dr. Ali Etemad
Research Areas: Continual Learning, Federated Learning, Pedestrian Detection
University of Saskatchewan
MSc in Electrical Engineering
May 2019 - Nov 2021 | GPA: 4.0/4.0Advisor: Dr. Seok-bum Ko
Thesis: "Deep Learning for Robust Super-Resolution"
Iran University of Science and Technology
BSc in Electrical Engineering - Electronics
Sep 2013 - Jan 2018 | GPA: 3.76/4.0Advisor: Dr. Shahriar B. Shokouhi
Thesis: "Implementation of a Recurrent Neural Network Using Memristor Crossbar Array"
Technical Skills
- Programming Languages: Python 3.x, MATLAB, C/C++, Verilog, VHDL, HSPICE
- Python Libraries: PyTorch, TensorFlow 2.x/Keras, Scikit-learn, Matplotlib/Seaborn, NumPy, SciPy, Pandas, OpenCV
- General Tools: Git, LaTeX, Anaconda, Excel, Adobe Photoshop
Honors & Awards
- Devolved Scholarship - Department of Electrical and Computer Engineering, USASK - $16,000 (2021)
- Best Poster Award - 4th Annual P2IRC Symposium, Saskatoon, Canada (2019)
- 90th Percentile - School of Electrical Engineering, IUST (2017)
Publications
For the full list of publications and citations, see my Google Scholar Profile.
Learning and Optimization Techniques
- Molahasani M, Greenspan M, Etemad A. "Continual Learning for Long-Tailed Recognition: Bridging the Gap in Theory and Practice." In NeurIPS Workshop on Mathematics of Modern Machine Learning (M3L), 2023.
- Molahasani M, Greenspan M, Etemad A. "Can Continual Learning Improve Long-Tailed Recognition? Toward a Unified Framework." arXiv preprint arXiv:2306.13275. 2023.
- Pourpanah F, Molahasani M, Soltany M, Greenspan M, Etemad A. "Federated Unsupervised Domain Generalization using Global and Local Alignment of Gradients." arXiv preprint arXiv:2405.16304. 2024.
AI in Medical Imaging and Health
- Haghanifar A, Molahasani M, Choi Y, Deivalakshmi S, Ko SB. "Covid-CXNet: Detecting Covid-19 in Frontal Chest X-ray Images Using Deep Learning." Multimedia Tools and Applications. 2022.
- Molahasani M, Choi Y, Deivalakshmi S, Ko SB. "Capsule GAN for Prostate MRI Super-resolution." Multimedia Tools and Applications. 2021.
- Haghanifar A, Molahasani M, Haghanifar S, Choi Y, Ko SB. "PaXNet: Tooth Segmentation and Dental Caries Detection in Panoramic X-ray Using Ensemble Transfer Learning and Capsule Classifier." Multimedia Tools and Applications. 2023.
- Haghanifar A, Molahasani M, Ko SB. "Automated Teeth Extraction From Dental Panoramic X-ray Images Using Genetic Algorithm." In IEEE International Symposium on Circuits and Systems (ISCAS) 2020.
Applied AI
- Molahasani M, Ko SB. "Capsule GAN for Robust Face Super Resolution." Multimedia Tools and Applications. 2020.
- Molahasani M, Etemad A, Greenspan M. "Out-of-Distribution Pedestrian Detection." In International Conference on Image Processing (ICIP) 2023.
- Zand M, Damirchi H, Farley A, Molahasani M, Greenspan M, Etemad A. "Multiscale Crowd Counting and Localization By Multitask Point Supervision." In International Conference on Acoustics, Speech, and Signal Processing (ICASSP) 2022.
- Etminan E, Molahasani M, Ko S, Wiens T. "Using Dynamic Pressure Response for Erosion Detection in Hydraulic Tubes and Hoses." In Fluid Power Systems Technology 2021.
- Molahasani M, Ko SB. "Msg-CapsGAN: Multi-scale Gradient Capsule GAN for Face Super Resolution." In International Conference on Electronics, Information, and Communication (ICEIC) 2020.
- Sharifi R, Molahasani M, Vakili VT. "Mobile User-Activity Prediction Utilizing LSTM Recurrent Neural Network." In IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM) 2019.
Implementation of ML
- Molahasani M, Shamsi J, Baradaran Shokouhi S. "Hybrid CMOS/Memristor Crossbar Structure for Implementing Hopfield Neural Network." Analog Integrated Circuits and Signal Processing. 2021.
- Molahasani M, Shokouhi SB, Ko SB. "Efficient Hybrid CMOS/Memristor Implementation of Bidirectional Associative Memory Using Passive Weight Array." Microelectronics Journal. 2020.
Teaching Experience
Introduction to Computer Programming for Engineers
Teaching Assistant - Queen's University (2022 - 2023)
Guided over 150 students in labs, enhancing their proficiency in C/C++ programming.
Computer Architecture
Teaching Assistant - Queen's University (2021 - 2022)
Assisted students in labs, facilitating hands-on experience in designing micro-computers using VHDL on FPGA boards.
Microprocessor Interfacing and Embedded Systems
Teaching Assistant - Queen's University (2020 - 2022)
Led lab sessions for 100 students, instructing them in assembly language programming on FPGA and employing C for I/O system integrations on connected extension boards.
Microprocessors
Teaching Assistant - Iran University of Science and Technology (2017 - 2018)
Tutored students, emphasizing I/O and interrupt handling techniques in Microprocessors.
Outreach & Professional Development
Peer Review
- IEEE Transactions on Multimedia 2024 (August 2024)
- NeurIPS 2024 (June 2024)
- ICML 2024 Workshop on Theoretical Foundations of Foundation Models (June 2024)
- AAAI 2024 Special Track on Safe, Robust, and Responsible Artificial Intelligence (Sep 2023)
- AAAI 2023 Workshop on Representation Learning for Responsible Human-Centric AI (Jan 2023)
- IEEE Transactions on Artificial Intelligence (Apr 2022, Feb 2022, Jan 2023)
- Machine Vision and Applications, Springer (May 2022)
- Imaging Science Journal, Taylor & Francis (Jan 2022, Jan 2023)
Service and Outreach
- VP of Student Affairs - Engineering Graduate Community Council at the USASK (EGCC) - 2020
- VP of Administration - Iranian Students’ Council at the University of Saskatchewan (ISC) - 2019
Competitions & Events
- WASP/Ingenuity Labs Poster Event - Presented work on continual pedestrian detection in the joint event between the Wallenberg AI, Autonomous Systems and Software Program (WASP) and Ingenuity Labs.
- Engineering Three Minute Thesis - Competed in the 2nd EGCC three-minute thesis (3MT) competition and organized the 3rd EGCC 3MT event.
References
- Michael Greenspan
Professor at the Department of Electrical and Computer Engineering, Queen's University
michael.greenspan@queensu.ca - Ali Etemad
Associate Professor at the Department of Electrical and Computer Engineering, Queen's University
ali.etemad@queensu.ca - Seok-bum Ko
Professor at the Department of Electrical and Computer Engineering and the Department of Biomedical Engineering, University of Saskatchewan
seokbum.ko@usask.ca - Shahriar B. Shokouhi
Associate Professor at School of Electrical Engineering, Iran University of Science and Technology
bshokouhi@iust.ac.ir