Picture for Zohar Ringel

Zohar Ringel

Applications of Statistical Field Theory in Deep Learning

Add code
Feb 25, 2025
Viaarxiv icon

From Kernels to Features: A Multi-Scale Adaptive Theory of Feature Learning

Add code
Feb 05, 2025
Viaarxiv icon

Symmetric Kernels with Non-Symmetric Data: A Data-Agnostic Learnability Bound

Add code
Jun 04, 2024
Viaarxiv icon

Wilsonian Renormalization of Neural Network Gaussian Processes

Add code
May 09, 2024
Viaarxiv icon

Towards Understanding Inductive Bias in Transformers: A View From Infinity

Add code
Feb 07, 2024
Viaarxiv icon

Droplets of Good Representations: Grokking as a First Order Phase Transition in Two Layer Networks

Add code
Oct 05, 2023
Figure 1 for Droplets of Good Representations: Grokking as a First Order Phase Transition in Two Layer Networks
Figure 2 for Droplets of Good Representations: Grokking as a First Order Phase Transition in Two Layer Networks
Figure 3 for Droplets of Good Representations: Grokking as a First Order Phase Transition in Two Layer Networks
Figure 4 for Droplets of Good Representations: Grokking as a First Order Phase Transition in Two Layer Networks
Viaarxiv icon

Speed Limits for Deep Learning

Add code
Jul 27, 2023
Viaarxiv icon

Spectral-Bias and Kernel-Task Alignment in Physically Informed Neural Networks

Add code
Jul 12, 2023
Viaarxiv icon

Separation of scales and a thermodynamic description of feature learning in some CNNs

Add code
Dec 31, 2021
Figure 1 for Separation of scales and a thermodynamic description of feature learning in some CNNs
Figure 2 for Separation of scales and a thermodynamic description of feature learning in some CNNs
Figure 3 for Separation of scales and a thermodynamic description of feature learning in some CNNs
Viaarxiv icon

A self consistent theory of Gaussian Processes captures feature learning effects in finite CNNs

Add code
Jun 08, 2021
Figure 1 for A self consistent theory of Gaussian Processes captures feature learning effects in finite CNNs
Figure 2 for A self consistent theory of Gaussian Processes captures feature learning effects in finite CNNs
Figure 3 for A self consistent theory of Gaussian Processes captures feature learning effects in finite CNNs
Figure 4 for A self consistent theory of Gaussian Processes captures feature learning effects in finite CNNs
Viaarxiv icon