Code for implementing the recursive metadata normalization layer, which removes the influence of confounding variables from intermediate learned features of a deep network.
Accompanying manuscript: https://openreview.net/pdf?id=7zrS5hHlfY
@inproceedings{
anonymous2025confounderfree,
title={Confounder-Free Continual Learning via Recursive Feature Normalization},
author={Yash Shah, Camila Gonzalez, MohammadHassan Abbasi, Qingyu Zhao, Kilian M. Pohl, Ehsan Adeli},
booktitle={Forty-second International Conference on Machine Learning},
year={2025},
url={https://openreview.net/pdf?id=7zrS5hHlfY}
}