3 citations · 8 across the 7 of their papers we have counts for
8 papers
Enhancing Neural Training via a Correlated Dynamics Model
Jonathan Brokman, Roy Betser, Rotem Turjeman +3
As neural networks grow in scale, their training becomes both computationally demanding and rich in dynamics. Amidst the flourishing interest in these training dynamics, we present…
A Minimal Set of Koopman Eigenfunctions -- Analysis and Numerics
Ido Cohen, Eli Appleboim
Research on Koopman operator theory has focused on three key areas for several decades: the mathematical structure of the Koopman eigenfunction space, the basis of this space, and…
The Underlying Correlated Dynamics in Neural Training
Rotem Turjeman, Tom Berkov, Ido Cohen +1
Training of neural networks is a computationally intensive task. The significance of understanding and modeling the training dynamics is growing as increasingly larger networks are…
BASiS: Batch Aligned Spectral Embedding Space
Or Streicher, Ido Cohen, Guy Gilboa
Graph is a highly generic and diverse representation, suitable for almost any data processing problem. Spectral graph theory has been shown to provide powerful algorithms, backed b…
Latent Modes of Nonlinear Flows -- a Koopman Theory Analysis
Ido Cohen, Guy Gilboa
Extracting the latent underlying structures of complex nonlinear local and nonlocal flows is essential for their analysis and modeling. In this work, we attempt to provide a consis…
Total-Variation -- Fast Gradient Flow and Relations to Koopman Theory
Ido Cohen, Tom Berkov, Guy Gilboa
The space-discrete Total Variation (TV) flow is analyzed using several mode decomposition techniques. In the one-dimensional case, we provide analytic formulations to Dynamic Mode…
Modes of Homogeneous Gradient Flows
Ido Cohen, Omri Azencot, Pavel Lifshitz +1
Finding latent structures in data is drawing increasing attention in diverse fields such as image and signal processing, fluid dynamics, and machine learning. In this work we exami…
Introducing the p-Laplacian Spectra
Ido Cohen, Guy Gilboa
In this work we develop a nonlinear decomposition, associated with nonlinear eigenfunctions of the p-Laplacian for p \in (1, 2). With this decomposition we can process signals of d…