About Me
Hi! This is Vahid! I am currently a Masters of Science in Applied Computing student at the CS Department of the University of Toronto. I enjoy researching both the theoretical and practical sides of Machine Learning and relish mathematically-inspired research in general.
Links: Github, LinkedIn, Google Scholar, Personal Email
Education
- Courses (CGPA: 4.0/4): Introduction to Causality (A+), Deep Learning: Theory & Datascience (A+), Neural Networks & Deep Learning, Empirical Software Engineering
- CGPA: 18.64/20 (3.8/4) — 141/140 credits
- Selected Courses (All Courses): Digital Image Processing (Graduate), Stochastic Processes (Math Dep.), Artificial Intelligence, Linear Algebra, Signals & Systems, Probability and Statistics, Design of Algorithms, Data Structures and Algorithms
- Selected Teaching Assistantships (All TAships): Head TA of Artificial Intelligence (Cross-appointed), supervised graduate TAs in Machine Learning Privacy (Graduate), tutored deep learning workshops in Machine Learning for Bioinformatics (Graduate), in addition to TAing and tutoring recitation classes for undergraduate courses such Bioinformatics, Probability theory and Statistics, Linear Algebra, and Data structures and Algorithms.
- Graduated top student, Diploma GPA: 19.68/20
Research Experience
- Studying the potential of diffusion processes and score-based models in stabilizing the learning of deep causal structures
- Co-developed a fully differentiable Bayesian approach for causal structure discovery and causal order discovery
- Developed DyPy & Lightning-Toolbox for fast implementation of complex deep learning research experiments
- Conducted a literature review on deep probabilistic model and deep generative models and deep anomaly detection
- Assessed the potential of various repurposed generative models in image anomaly detection
- Developed a novel deep training procedure for anomaly detection using explicit density estimation while learning a latent space
- Developed TorchDE, a unified pytorch library for deep density estimation
École Polytechnique Fédérale de Lausanne (EPFL)
Co-supervised by the LTS4 & MLO Labs
Feb. 2021 - Sep. 2021
- Conducted a literature review on the theoretical foundations of deep learning and non-convolutional vision neural networks
- Working under two PhD students, studied the directional inductive bias of vision transformer network architectures
- Assessed the state-of-the-art vehicle trajectory prediction models’ explainability and reliance on different scene-data priors
- Explored NLP methods to embrace input sparsity using generic SVG-based inputs in vehicle trajectory prediction models
Professional Experience
- Designed and developed a novel satellite and aerial imagery super-resolution model for urban maps
- Designed and implemented a generic composable deep learning training and prototyping framework
- Implemented a Kubernetes cluster node failure simulator and designed a cluster monitoring system for internal experiments
- Designed and implemented a novel decentralized load-tester based on the company's true user-behavior data
Pido
Computer Vision Engineer
Jun. 2018 - Jun. 2019
- Implemented an efficiently accurate solution for extracting bilingual (Farsi and English) information from Iranian debit card scans
Publications
Anomaly Detection via Explicit Density Estimation (Bachelor's thesis)
SVG-Net: An SVG-Based Trajectory Prediction Model
Bahari, M.,
Zehtab, V.*,
Khorasani, S.*,
Ayromlou, S.*,
Saadatnejad, S.,
Alahi, A.
—*Equal contribution as 2nd Author