Itay Safran

Most-cited papers

Spurious Local Minima are Common in Two-Layer ReLU Neural Networks
Itay Safran, Ohad Shamir
2017International Conference on Machine Learning295 citations
Semantic Scholar
Depth-Width Tradeoffs in Approximating Natural Functions with Neural Networks
Itay Safran, Ohad Shamir
2016International Conference on Machine Learning189 citations
Semantic Scholar
On the Quality of the Initial Basin in Overspecified Neural Networks
Itay Safran, Ohad Shamir
2015International Conference on Machine Learning131 citations
Semantic Scholar
How Good is SGD with Random Shuffling?
Itay Safran, Ohad Shamir
2019Annual Conference Computational Learning Theory105 citations
Semantic Scholar
A Simple Explanation for the Existence of Adversarial Examples with Small Hamming Distance
A. Shamir, Itay Safran, Eyal Ronen, O. Dunkelman
2019arXiv.org97 citations
Semantic Scholar
The Effects of Mild Over-parameterization on the Optimization Landscape of Shallow ReLU Neural Networks
Itay Safran, Gilad Yehudai, Ohad Shamir
2020Annual Conference Computational Learning Theory46 citations
Semantic Scholar
Depth Separations in Neural Networks: What is Actually Being Separated?
Itay Safran, Ronen Eldan, Ohad Shamir
2019Constructive approximation43 citations Open access
Semantic Scholar
Depth Separation in ReLU Networks for Approximating Smooth Non-Linear Functions
Itay Safran, Ohad Shamir
2016arXiv.org32 citations
Semantic Scholar
On the Effective Number of Linear Regions in Shallow Univariate ReLU Networks: Convergence Guarantees and Implicit Bias
Itay Safran, Gal Vardi, Jason D. Lee
2022Neural Information Processing Systems27 citations Open access
Semantic Scholar
Random Shuffling Beats SGD Only After Many Epochs on Ill-Conditioned Problems
Itay Safran, Ohad Shamir
2021Neural Information Processing Systems26 citations
Semantic Scholar
Optimization-Based Separations for Neural Networks
Itay Safran, Jason D. Lee
2021Annual Conference Computational Learning Theory22 citations
Semantic Scholar
How Many Neurons Does it Take to Approximate the Maximum?
Itay Safran, Daniel Reichman, Paul Valiant
2023ACM-SIAM Symposium on Discrete Algorithms13 citations Open access
Semantic Scholar
To Grok Grokking: Provable Grokking in Ridge Regression
Mingyue Xu, Gal Vardi, Itay Safran
2026arXiv.org6 citations
Semantic Scholar
A Depth Hierarchy for Computing the Maximum in ReLU Networks via Extremal Graph Theory
Itay Safran
2026Annual Conference Computational Learning Theory4 citations
Semantic Scholar
Provable Privacy Attacks on Trained Shallow Neural Networks
Guy Smorodinsky, Gal Vardi, Itay Safran
2024arXiv.org4 citations
Semantic Scholar
Depth Separations in Neural Networks: Separating the Dimension from the Accuracy
Itay Safran, Daniel Reichman, Paul Valiant
2024Annual Conference Computational Learning Theory4 citations
Semantic Scholar
No Prior, No Leakage: Revisiting Reconstruction Attacks in Trained Neural Networks
Yehonathan Refael, Guy Smorodinsky, Ofir Lindenbaum, Itay Safran
2025arXiv.org2 citations
Semantic Scholar
The Median is Easier than it Looks: Approximation with a Constant-Depth, Linear-Width ReLU Network
Abhigyan Dutta, Itay Safran, Paul Valiant
2026Annual Conference Computational Learning Theory1 citations
Semantic Scholar
On the Rate of Convergence of GD in Non-linear Neural Networks: An Adversarial Robustness Perspective
Guy Smorodinsky, Sveta Gimpleson, Itay Safran
2026arXiv.org0 citations
Semantic Scholar
Every Layer Counts: An Exponential $L_2$ Depth Hierarchy for ReLU Networks
Itay Safran
20260 citations
Semantic Scholar