Thomas George
@tfjgeorge.bsky.social
Explainability of deep neural nets and causality https://tfjgeorge.github.io/
Very interesting challenge! How will you balance accuracy and energy efficiency in your final score?
January 9, 2025 at 3:49 PM
Very interesting challenge! How will you balance accuracy and energy efficiency in your final score?
Congratulations for a very interesting paper! On the same topic, allow me to adverstise our AISTATS paper arxiv.org/abs/2008.00938 where we use the "sum of linearized steps" view to derive a Rademacher complexity bound which uses tangent features during training (fig. 6).
Implicit Regularization via Neural Feature Alignment
We approach the problem of implicit regularization in deep learning from a geometrical viewpoint. We highlight a regularization effect induced by a dynamical alignment of the neural tangent features i...
arxiv.org
November 21, 2024 at 3:46 PM
Congratulations for a very interesting paper! On the same topic, allow me to adverstise our AISTATS paper arxiv.org/abs/2008.00938 where we use the "sum of linearized steps" view to derive a Rademacher complexity bound which uses tangent features during training (fig. 6).