Disentangling Identifiable Features from Noisy Data with Structured Nonlinear ICA

This repository contains code for our $\Delta$-SNICA model from the paper, published at NeurIPS 2021. If you use our code/model for your research, we would be grateful if you cite our work as:

@inproceedings{HalvaSNICA21,
 author = {H\"{a}lv\"{a}, Hermanni and Le Corff, Sylvain and Leh\'{e}ricy, Luc and So, Jonathan and Zhu, Yongjie and Gassiat, Elisabeth and Hyvarinen, Aapo},
 booktitle = {Advances in Neural Information Processing Systems},
 editor = {M. Ranzato and A. Beygelzimer and Y. Dauphin and P.S. Liang and J. Wortman Vaughan},
 pages = {1624--1633},
 publisher = {Curran Associates, Inc.},
 title = {Disentangling Identifiable Features from Noisy Data with Structured Nonlinear ICA},
 url = {https://proceedings.neurips.cc/paper_files/paper/2021/file/0cdbb4e65815fbaf79689b15482e7575-Paper.pdf},
 volume = {34},
 year = {2021}

License

A full copy of the license can be found here