today local_bar

Andres Guzman-Cordero

PhD student, Université de Montréal

1 paper at NeurIPS 2025

Homepage· OpenReview· Semantic Scholar· Google Scholar

Poster Session 4

Thursday, December 4, 2025 · 4:30 PM → 7:30 PM

Exhibit Hall C,D,E

Improving Energy Natural Gradient Descent through Woodbury, Momentum, and Randomization

#2209 · Andres Guzman-Cordero, Felix Dangel, Gil Goldshlager, Marius Zeinhofer

We introduce Woodbury's matrix identity, momentum-like SPRING and randomization to make energy natural gradient descent 75 times faster for PINNs.