A Unified Approach to Feature Learning in Bayesian Neural Networks
Abstract
The power of neuronal networks comes from their adaptation to training data, known as feature learning. We consider feature learning within Bayesian learning and derive the two prominent high dimensional theories, kernel scaling and kernel adaptation, respectively, from a unified large deviation approach. We then show when feature learning escapes the scaling approach, but is captured by kernel adaptation.
Cite
Text
Rubin et al. "A Unified Approach to Feature Learning in Bayesian Neural Networks." ICML 2024 Workshops: HiLD, 2024.Markdown
[Rubin et al. "A Unified Approach to Feature Learning in Bayesian Neural Networks." ICML 2024 Workshops: HiLD, 2024.](https://mlanthology.org/icmlw/2024/rubin2024icmlw-unified/)BibTeX
@inproceedings{rubin2024icmlw-unified,
title = {{A Unified Approach to Feature Learning in Bayesian Neural Networks}},
author = {Rubin, Noa and Ringel, Zohar and Seroussi, Inbar and Helias, Moritz},
booktitle = {ICML 2024 Workshops: HiLD},
year = {2024},
url = {https://mlanthology.org/icmlw/2024/rubin2024icmlw-unified/}
}