Functional Gradient Descent Framework Accepted to NeurIPS, Outperforming Neural Networks

The paper 'Functional Gradient Descent with Adaptive Representations' has been accepted to NeurIPS, tackling long-standing challenges in functional gradient descent. Experiments show the framework outperforms neural networks by an order of magnitude, with the first author answering questions from the community.

2026-09-27 ~ 2026-09-28 · 2 related posts

• Functional Gradient Descent with Adaptive Representations accepted at NeurIPS — CatAstro_Piyush · 2026-09-27
• Functional Gradient Descent with Adaptive Representations accepted at NeurIPS, beats neural nets by ~10x — dccsillag0 · 2026-09-28