Nadav Cohen  —  Publications (see also Google Scholar)

+ Supervised student paper
* Primary authorship

Preprints

+ Implicit Bias of Policy Gradient in Linear Quadratic Control: Extrapolation to Unseen Initial States
Noam Razin, Yotam Alexander, Edo Cohen-Karlik, Raja Giryes, Amir Globerson and Nadav Cohen. Feb’24.
Preprint.

Conference Proceedings

+ What Makes Data Suitable for a Locally Connected Neural Network? A Necessary and Sufficient Condition Based on Quantum Entanglement (extended arXiv version)
Yotam Alexander, Nimrod De La Vega, Noam Razin and Nadav Cohen. Mar’23 (v1), Oct’23 (v2).
Conference on Neural Information Processing Systems (NeurIPS) 2023, Spotlight Track (top 3%).

+ On the Ability of Graph Neural Networks to Model Interactions Between Vertices (extended arXiv version)
Noam Razin, Tom Verbin and Nadav Cohen. Nov’22 (v1), Oct’23 (v2).
Conference on Neural Information Processing Systems (NeurIPS) 2023.

+ Learning Low Dimensional State Spaces with Overparameterized Recurrent Neural Nets
Edo Cohen-Karlik, Itamar Menuhin-Gruman, Raja Giryes, Nadav Cohen and Amir Globerson. Oct’22 (v1), Mar’23 (v2).
International Conference on Learning Representations (ICLR) 2023.

+ Implicit Regularization in Hierarchical Tensor Factorization and Deep Convolutional Neural Networks (extended arXiv version)
Noam Razin, Asaf Maman and Nadav Cohen. Jan’22.
International Conference on Machine Learning (ICML) 2022.

On the Implicit Bias of Gradient Descent for Temporal Extrapolation
Edo Cohen-Karlik, Avichai Ben David, Nadav Cohen and Amir Globerson. Feb’22.
Conference on Artificial Intelligence and Statistics (AISTATS) 2022.

+ Continuous vs. Discrete Optimization of Deep Neural Networks (extended arXiv version)
Omer Elkabetz and Nadav Cohen. Jul’21 (v1), Dec’21 (v2).
Conference on Neural Information Processing Systems (NeurIPS) 2021, Spotlight Track (top 3%).

+ Implicit Regularization in Tensor Factorization
Noam Razin, Asaf Maman and Nadav Cohen. Feb’21.
International Conference on Machine Learning (ICML) 2021.

+ Implicit Regularization in Deep Learning May Not Be Explainable by Norms (extended arXiv version)
Noam Razin and Nadav Cohen. May’20 (v1), Oct’20 (v2).
Conference on Neural Information Processing Systems (NeurIPS) 2020.

* Implicit Regularization in Deep Matrix Factorization
Sanjeev Arora, Nadav Cohen, Wei Hu and Yuping Luo (alphabetical order). Jun’19 (v1), Oct’19 (v2).
Conference on Neural Information Processing Systems (NeurIPS) 2019, Spotlight Track (top 3%).

* A Convergence Analysis of Gradient Descent for Deep Linear Neural Networks
Sanjeev Arora, Nadav Cohen, Noah Golowich and Wei Hu (alphabetical order). Oct’18 (v1), Nov’18 (v2).
International Conference on Learning Representations (ICLR) 2019.

* On the Optimization of Deep Networks: Implicit Acceleration by Overparameterization
Sanjeev Arora, Nadav Cohen and Elad Hazan (alphabetical order). Feb’18.
International Conference on Machine Learning (ICML) 2018.

“Zero-Shot” Super-Resolution Using Deep Internal Learning
Assaf Shocher, Nadav Cohen and Michal Irani. Dec’17.
Conference on Computer Vision and Pattern Recognition (CVPR) 2018.

* Boosting Dilated Convolutional Networks with Mixed Tensor Decompositions (extended arXiv version)
Nadav Cohen, Ronen Tamari and Amnon Shashua. Apr’17.
International Conference on Learning Representations (ICLR) 2018, Oral Track (top 1%).

Deep Learning and Quantum Entanglement: Fundamental Connections with Implications to Network Design (extended arXiv version)
Yoav Levine, David Yakira, Nadav Cohen and Amnon Shashua. Apr’17.
International Conference on Learning Representations (ICLR) 2018.

* Inductive Bias of Deep Convolutional Networks through Pooling Geometry
Nadav Cohen and Amnon Shashua. May’16 (v1), Nov’16 (v2).
International Conference on Learning Representations (ICLR) 2017.

* Convolutional Rectifier Networks as Generalized Tensor Decompositions (extended arXiv version)
Nadav Cohen and Amnon Shashua. Mar’16.
International Conference on Machine Learning (ICML) 2016.

* On the Expressive Power of Deep Learning: A Tensor Analysis
Nadav Cohen, Or Sharir and Amnon Shashua. Sep’15 (v1), Feb’16 (v2).
Conference on Learning Theory (COLT) 2016.

* Deep SimNets
Nadav Cohen, Or Sharir and Amnon Shashua. Jun’15 (v1), Nov’15 (v2).
Conference on Computer Vision and Pattern Recognition (CVPR) 2016.

Journals

Deep Linear Networks for Matrix Completion  —  An Infinite Depth Limit
Nadav Cohen, Govind Menon and Zsolt Veraszto. Oct’22.
Society for Industrial and Applied Mathematics (SIAM) Journal on Applied Dynamical Systems.

Quantum Entanglement in Deep Learning Architectures (arXiv version)
Yoav Levine, Or Sharir, Nadav Cohen, Amnon Shashua. Mar’18 (v1), Feb’19 (v2).
Physical Review Letters (PRL).

Book Chapters

Bridging Many-Body Quantum Physics and Deep Learning via Tensor Networks
Yoav Levine, Or Sharir, Nadav Cohen, Amnon Shashua. Dec’22.
Mathematical Aspects of Deep Learning, Cambridge University Press.

Tensors for Deep Learning Theory: Analyzing Deep Learning Architectures via Tensorization
Yoav Levine, Noam Wies, Or Sharir, Nadav Cohen, Amnon Shashua. Nov’21.
Tensors for Data Processing: Theory, Methods and Applications, Academic Press.

Invited Papers, Workshops and Technical Reports

* Analysis and Design of Convolutional Networks via Hierarchical Tensor Decompositions
Nadav Cohen, Or Sharir, Yoav Levine, Ronen Tamari, David Yakira and Amnon Shashua. Jun’17.
Intel Collaborative Research Institute Special Issue on Deep Learning Theory.

Tensorial Mixture Models
Or Sharir, Ronen Tamari, Nadav Cohen and Amnon Shashua. Oct’16.
Technical Report.

* SimNets: A Generalization of Convolutional Networks
Nadav Cohen and Amnon Shashua. Oct’14 (v1), Dec’14 (v2).
Conference on Neural Information Processing Systems (NeurIPS) 2014, Deep Learning Workshop.