About Me
I am a Postdoctoral Fellow in Applied Mathematics and Computer Science at Harvard University,
working with Prof. Melanie Weber.
I completed my Ph.D. at MIT under the supervision of Prof. Stefanie Jegelka.
My research interests lie at the intersection of
Geometric Machine Learning (symmetries, manifolds, graphs),
Deep Learning Theory, and the
foundations of Large Language Models (LLMs).
More recently, from an applied mathematics perspective, I have been particularly interested in
applied group representation theory,
harmonic (Fourier) analysis,
spectral theory of manifolds, and
differential geometry,
and in exploring their connections to machine learning and statistics.
For more about my research, please visit my recent tutorial at NeurIPS 2025 in San Diego, CA
(click here).
Published Papers
Note: Authors denoted by * contributed equally.
• Achieving Approximate Symmetry Is Exponentially Easier than Exact Symmetry
[arXiv]
Behrooz Tahmasebi, Melanie Weber
ICLR 2026
• Geometric Algorithms for Neural Combinatorial Optimization with Constraints
[arXiv]
Nikolaos Karalias, Akbar Rafiey, Yifei Xu, Zhishang Luo, Behrooz Tahmasebi, Connie Jiang, Stefanie Jegelka
NeurIPS 2025
• Learning with Exact Invariances in Polynomial Time
[preprint]
Ashkan Soleymani*, Behrooz Tahmasebi*, Stefanie Jegelka, Patrick Jaillet
ICML 2025, Spotlight (top 2.6% of submissions)
• Generalization Bounds for Canonicalization: A Comparative Study with Group Averaging
[conference-link]
[link]
Behrooz Tahmasebi, Stefanie Jegelka
ICLR 2025
• A Robust Kernel Statistical Test of Invariance: Detecting Subtle Asymmetries
[link]
Ashkan Soleymani*, Behrooz Tahmasebi*, Stefanie Jegelka, Patrick Jaillet
AISTATS 2025, Oral (top 2% of submissions)
• Regularity in Canonicalized Models: A Theoretical Perspective
[link]
Behrooz Tahmasebi, Stefanie Jegelka
AISTATS 2025
• Coded Computing for Resilient Distributed Computing: A Learning-Theoretic Framework
[arXiv]
Parsa Moradi, Behrooz Tahmasebi, Mohammad Ali Maddah-Ali
NeurIPS 2024
• A Universal Class of Sharpness-Aware Minimization Algorithms
[arXiv]
[conference-link]
[workshop-link]
Behrooz Tahmasebi, Ashkan Soleymani, Dara Bahri, Stefanie Jegelka, Patrick Jaillet
ICML 2024
Best Paper Award, workshop on High-dimensional Learning Dynamics (HiLD) at ICML 2024
• Sample Complexity Bounds for Estimating Probability Divergences under Invariances
[arXiv]
Behrooz Tahmasebi, Stefanie Jegelka
ICML 2024
• The Exact Sample Complexity Gain from Invariances for Kernel Regression
[arXiv]
Behrooz Tahmasebi, Stefanie Jegelka
NeurIPS 2023, Spotlight (top 3.6% of submissions)
• The Power of Recursion in Graph Neural Networks for Counting Substructures
[conference-link]
Behrooz Tahmasebi, Derek Lim, Stefanie Jegelka
AISTATS 2023, Oral (top 1.9% of submissions)
• The Capacity of Associated Subsequence Retrieval
[journal-link]
Behrooz Tahmasebi, Mohammad Ali Maddah-Ali, Seyed Abolfazl Motahari
IEEE Transactions on Information Theory, 2021
• Private Function Computation
Behrooz Tahmasebi, Mohammad Ali Maddah-Ali
IEEE International Symposium on Information Theory (ISIT), 2020
• Private Sequential Function Computation
[arXiv]
Behrooz Tahmasebi, Mohammad Ali Maddah-Ali
IEEE International Symposium on Information Theory (ISIT), 2019
• Information Theory of Mixed Population Genome-Wide Association Studies
Behrooz Tahmasebi, Mohammad Ali Maddah-Ali, Seyed Abolfazl Motahari
IEEE Information Theory Workshop (ITW), 2018
• Genome-Wide Association Studies: Information Theoretic Limits of Reliable Learning
Behrooz Tahmasebi, Mohammad Ali Maddah-Ali, Seyed Abolfazl Motahari
IEEE International Symposium on Information Theory (ISIT), 2018
• Optimum Transmission Delay for Function Computation in NFV-Based Networks: The Role of Network Coding and Redundant Computing
Behrooz Tahmasebi, Mohammad Ali Maddah-Ali, Saeedeh Parsaeefard, Babak Khalaj
IEEE Journal on Selected Areas in Communications (JSAC), 2018
• On the Identifiability of Finite Mixtures of Finite Product Measures
Behrooz Tahmasebi, Seyed Abolfazl Motahari, Mohammad Ali Maddah-Ali
Manuscript, 2018
