Omar Montasser

Omar Montasser

  Google Scholar





News: I have joined UC Berkeley as a FODSI-Simons postdoctoral researcher. Following that, in July 2024, I will start as an Assistant Professor at Yale, Department of Statistics and Data Science!

My research broadly explores the theory of machine learning. I recently completed my PhD at the Toyota Technological Institute at Chicago, where I was fortunate to be advised by Nathan Srebro. During my PhD, I primarily thought about and worked on questions related to learning machine learning models robust against adversarial examples, through the lens of learning theory. For a broad overview, you can watch my PhD thesis defense, and/or read my PhD thesis for an in-depth exploration.

Before TTIC, I completed a five-year program (combined BS/MS) in computer science and engineering at Penn State. While there, I enjoyed working with Daniel Kifer and Sean Hallgren on problems in machine learning and quantum computational complexity.


Adversarially Robust Learning: A Generic Minimax Optimal Learner & Characterization
Omar Montasser, Steve Hanneke, and Nathan Srebro
NeurIPS, 2022. (Oral)

Boosting Barely Robust Learners: A New Perspective on Adversarial Robustness
Avrim Blum, Omar Montasser, Greg Shakhnarovich, Hongyang Zhang
NeurIPS, 2022.

A Theory of PAC Learnability under Transformation Invariances
Han Shao, Omar Montasser, and Avrim Blum
NeurIPS, 2022. (Oral)

Transductive Robust Learning Guarantees
Omar Montasser, Steve Hanneke, and Nathan Srebro
AISTATS, 2022.

Adversarially Robust Learning with Unknown Perturbation Sets
Omar Montasser, Steve Hanneke, and Nathan Srebro
COLT, 2021.

Beyond Perturbations: Learning Guarantees with Arbitrary Adversarial Test Examples
Shafi Goldwasser, Adam Tauman Kalai, Yael Tauman Kalai, Omar Montasser
NeurIPS, 2020. (Spotlight)

Reducing Adversarially Robust Learning to Non-Robust PAC Learning
Omar Montasser, Steve Hanneke, and Nathan Srebro
NeurIPS, 2020.

Efficiently Learning Adversarially Robust Halfspaces with Noise
Omar Montasser, Subhi Goel, Ilias Diakonikolas, and Nathan Srebro
ICML, 2020.

Approximate is Good Enough: Probabilistic Variants of Dimensional and Margin Complexity
Pritish Kamath, Omar Montasser, and Nathan Srebro
COLT, 2020.

VC Classes are Adversarially Robustly Learnable, but Only Improperly
Omar Montasser, Steve Hanneke, and Nathan Srebro
COLT, 2019. Best Student Paper Award!

Predicting Demographics of High-Resolution Geographies with Geotagged Tweets
Omar Montasser and Daniel Kifer
AAAI, 2017. (Oral)


Based on minimal theme by orderedlist.