Behrooz Tahmasebi
Ph.D. student in EECS at MIT
Student member of CSAIL
Advisor: Stefanie Jegelka
CV |
Google Scholar |
Email: bzt at mit dot edu
About Me
I'm interested in learning with graphs, manifolds, and invariances, deep learning (geometry, optimization), and foundations of Large Language Models (LLMs). Previously, I studied at Sharif University of Technology.
Published Papers
Note: Authors denoted by * contributed equally.
• Learning with Exact Invariances in Polynomial Time
• Generalization Bounds for Canonicalization: A Comparative Study with Group Averaging
• A Robust Kernel Statistical Test of Invariance: Detecting Subtle Asymmetries
• Regularity in Canonicalized Models: A Theoretical Perspective
• Coded Computing for Resilient Distributed Computing: A Learning-Theoretic Framework
• A Universal Class of Sharpness-Aware Minimization Algorithms
• Sample Complexity Bounds for Estimating Probability Divergences under Invariances
• The Exact Sample Complexity Gain from Invariances for Kernel Regression
• The Power of Recursion in Graph Neural Networks for Counting Substructures
• The Capacity of Associated Subsequence Retrieval
• Private Function Computation
• Private Sequential Function Computation
• Information Theory of Mixed Population Genome-Wide Association Studies
• Genome-Wide Association Studies: Information Theoretic Limits of Reliable Learning
• Optimum Transmission Delay for Function Computation in NFV-Based Networks: The Role of Network Coding and Redundant Computing
• On the Identifiability of Finite Mixtures of Finite Product Measures