Neha S. Wadia, Ryan Zarcone, Michael R. DeWeese
A Solution to the Fokker-Planck Equation for Slowly Driven Brownian Motion: Emergent Geometry and a Formula for the Corresponding Thermodynamic Metric
Physical Review E, 2022. journal arxiv
Neha S. Wadia, Daniel Duckworth, Sam Schoenholz, Ethan Dyer, Jascha Sohl-Dickstein
Whitening and Second Order Optimization Both Make Information in the Dataset Unusable During Training, and Can Reduce or Prevent Generalization
Proceedings of the 27th International Conference on Machine Learning (ICML), 2021. journal arxiv
Charles Frye, Jamie Simon, Neha S. Wadia, Andrew Ligeralde, Michael R. DeWeese, Kristofer Bouchard
Critical Point-Finding Methods Reveal Gradient-Flat Regions of Deep Network Losses
Neural Computation, 2021. journal arxiv
Eric B. Norrgard, Eustace R. Edwards, Daniel J. McCarron, Matthew H. Steinecker, David DeMille, Shah Saad Alam, Stephen K. Peck, Neha S. Wadia, Larry R. Hunter
Hyperfine structure of the $B^3\Pi_1$ state and predictions of optical cycling behavior in the $X\rightarrow B$ transition of TlF
Physical Review A, 2017. journal arxiv
Neha S. Wadia, Michael I. Jordan, Michael Muehlebach
Optimization with Adaptive Step Size Selection from a Dynamical Systems Perspective
NeurIPS Workshop on Optimization for Machine Learning (OptML), 2021. link
Neha S. Wadia, Yatin Dandi, Michael I. Jordan
A Gentle Introduction to Gradient-Based Optimization and Variational Inequalities for Machine Learning
In review at J. Stat. Mech. (JSTAT). arxiv
Charles Frye, Neha S. Wadia, Michael R. DeWeese, Kristofer Bouchard
Numerically Recovering the Critical Points of a Deep Linear Autoencoder
arXiv preprint, 2019. arxiv
Slowly Driven Brownian Motion: Emergent Geometry and the Thermodynamic Metric
The Impact of Whitening and Second-Order Optimization on Generalization