Skip to content

Latest commit

 

History

History
45 lines (32 loc) · 1.72 KB

README.md

File metadata and controls

45 lines (32 loc) · 1.72 KB

Fair Streaming PCA

This is an official repository containing codes for our Fair Streaming Principal Component Analysis: Statistical and Algorithmic Viewpoint, accepted at NeurIPS 2023.

CelebA Dataset (celeba_fair_streaming_pca/)

Preparation

  1. Download CelebA dataset images (img_align_celeba.zip) from this link
  2. Download CelebA dataset annotations (list_attr_celeba.txt, ...) from this link
  3. Put these into the directory celeba_fair_streaming_pca/datasets/celeba/. Unzip the zip file here.

Running Codes

You may open the following four notebook files to run by yourself:

FairStreamingPCA_CelebA_RGB.ipynb
FairStreamingPCA_CelebA_grayscale.ipynb
FairStreamingPCA_CelebA_blocksizeAblation.ipynb
FairStreamingPCA_CelebA_rankAblation.ipynb

UCI Dataset (downstream_tasks_fair_streaming_pca/)

We mostly follow the instructions in this repository.

Synthetic Experiments (synthetic_tasks_fair_streaming_pca/)

Citation

If you'd like to use our code and publish material, please cite our paper:

@inproceedings{
    lee2023fair,
    title={Fair Streaming Principal Component Analysis: Statistical and Algorithmic Viewpoint},
    author={Junghyun Lee and Hanseul Cho and Se-Young Yun and Chulhee Yun},
    booktitle={Advances in Neural Information Processing Systems},
    year={2023},
    volume={36},
    pages={5126--5167},
    url={https://openreview.net/forum?id=TW3ipYdDQG}
}