Personal profile
Research interests
My research focuses on theoretical machine learning, with emphasis on deep learning theory. In particular, I am interested in the intersection of deep learning theory with related fields, such as approximation theory, optimization and theoretical computer science.
Fingerprint
- 1 Similar Profiles
Collaborations and top research areas from the last five years
-
Depth Separations in Neural Networks: Separating the Dimension from the Accuracy
Safran, I., Reichman, D. & Valiant, P., 1 Jan 2025, In: Proceedings of Machine Learning Research. 291Research output: Contribution to journal › Conference article › peer-review
-
How Many Neurons Does it Take to Approximate the Maximum?
Safran, I., Reichman, D. & Valiant, P., 1 Jan 2024, p. 3156-3183. 28 p.Research output: Contribution to conference › Paper › peer-review
6 Scopus citations -
Depth Separations in Neural Networks: What is Actually Being Separated?
Safran, I., Eldan, R. & Shamir, O., 1 Feb 2022, In: Constructive Approximation. 55, 1, p. 225-257 33 p.Research output: Contribution to journal › Article › peer-review
Open Access4 Scopus citations -
On the Effective Number of Linear Regions in Shallow Univariate ReLU Networks: Convergence Guarantees and Implicit Bias
Safran, I., Vardi, G. & Lee, J. D., 1 Jan 2022, Advances in Neural Information Processing Systems 35 - 36th Conference on Neural Information Processing Systems, NeurIPS 2022. Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K. & Oh, A. (eds.). Neural information processing systems foundation, (Advances in Neural Information Processing Systems; vol. 35).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review
15 Scopus citations -
Optimization-Based Separations for Neural Networks
Safran, I. & Lee, J. D., 1 Jan 2022, In: Proceedings of Machine Learning Research. 178, p. 3-64 62 p.Research output: Contribution to journal › Conference article › peer-review
9 Scopus citations -
Random Shuffling Beats SGD Only After Many Epochs on Ill-Conditioned Problems
Safran, I. & Shamir, O., 1 Jan 2021, Advances in Neural Information Processing Systems 34 - 35th Conference on Neural Information Processing Systems, NeurIPS 2021. Ranzato, M., Beygelzimer, A., Dauphin, Y., Liang, P. S. & Wortman Vaughan, J. (eds.). Neural information processing systems foundation, p. 15151-15161 11 p. (Advances in Neural Information Processing Systems; vol. 18).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review
14 Scopus citations -
The Effects of Mild Over-parameterization on the Optimization Landscape of Shallow ReLU Neural Networks
Safran, I., Yehudai, G. & Shamir, O., 1 Jan 2021, In: Proceedings of Machine Learning Research. 134, p. 3889-3934 46 p.Research output: Contribution to journal › Conference article › peer-review
15 Scopus citations -
How Good is SGD with Random Shuffling?
Safran, I. & Shamir, O., 1 Jan 2020, In: Proceedings of Machine Learning Research. 125, p. 3250-3284 35 p.Research output: Contribution to journal › Conference article › peer-review
56 Scopus citations -
Depth Separations in Neural Networks: What is Actually Being Separated?
Safran, I., Eldan, R. & Shamir, O., 1 Jan 2019, In: Proceedings of Machine Learning Research. 99, p. 2664-2666 3 p.Research output: Contribution to journal › Conference article › peer-review
19 Scopus citations -
Spurious local minima are common in two-layer ReLU neural networks
Safran, I. & Shamir, O., 1 Jan 2018, 35th International Conference on Machine Learning, ICML 2018. Krause, A. & Dy, J. (eds.). International Machine Learning Society (IMLS), p. 7031-7052 22 p. (35th International Conference on Machine Learning, ICML 2018; vol. 10).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review
40 Scopus citations
Projects
- 1 Active
-
Towards a Depth Hierarchy in Artificial Neural Networks
Safran, I. (PI)
1/01/25 → 31/12/28
Project: Research