Deep Learning Weight Pruning with RMT-SVD: Increasing Accuracy and Reducing Overfitting
Published:
This paper applies random matrix theory techniques to prune the weight matrices of fully connected neural networks. Arxiv Preprint
Download here
Publications
Complex Hanner’s Inequality for Many Functions
Published:
This paper comprises the results of my master’s thesis, edited for publication. To appear for publication at Israel Journal of Mathematics.
Download here
Surfactant Dynamics from the Arnold Perspective
Published:
This paper derives multiple partial differential equations that each govern fluid flow with additional surfactant dynamics. Techniques used include variational calculus. Published in Siam’s SIURO.
Download here
Central limit theorems for compound paths on the 2-dimensional lattice
Published:
This paper uses probability and counting techniques to derive central limit type theorems for a particular construction that is a higher dimensional analog of Zeckendorf’s decomposition. Published in The Fibonacci Quarterly.
Download here