Skip to content

README.md

Latest commit

 

History

History
75 lines (68 loc) · 2.54 KB

README.md

File metadata and controls

75 lines (68 loc) · 2.54 KB

Stochastic Momentum Methods for Non-smooth Non-Convex Finite-Sum Coupled Compositional Optimization

This repository provide codes for the paper Stochastic Momentum Methods for Non-smooth Non-Convex Finite-Sum Coupled Compositional Optimization .

Installation

# Clone the repository
git clone https://github.com/charlesxyc358/Stoc_mmt_NSNC_FCCO
cd Stoc_mmt_NSNC_FCCO
conda create -n fcco python==3.11
conda activate fcco
pip install -r requirements.txt

GroupDRO

First navigate to GroupDRO folder

cd GroupDRO

Download and Prepare the (labeled)Camelyon17_v1.0 and Amazon_v2.1 datasets

Download from Wilds Datasets, then put the downloaded tar.gz files under ./data folder and extract to the corresponding folder:

mkdir data
mkdir data/camelyon17_v1.0
tar -xzf ./data/camelyon17_v1.0.tar.gz -C ./data/camelyon17_v1.0
mkdir data/amazon_v2.1
tar -xzf ./data/amazon_v2.1.tar.gz -C ./data/amazon_v2.1

Train on Camelyon17 dataset with SONEX

python main_came.py --alg SONEX --gpu 0 --lr 1e-3 --beta 0.1 --gamma 0.2 --lamda 0.1 --theta 0.05 \
    --lr_c 0.1 --seed 1 --epoch 10 --alpha 0.15

Train on Amazon dataset with SONEX

python main_amz.py --alg SONEX --gpu 0 --lr 2e-5 --beta 0.1 --gamma 0.2 --lamda 0.1 --theta 0.01 \
    --wd 1e-2 --lr_c 0.1 --seed 1 --epoch 4 --alpha 0.15

AUC Maximization with ROC Fairness Constraints

First navigate to AUCMax folder

cd AUCMax

Download and Prepare the adult and compas datasets

Run the following command to download the datasets.

mkdir data
mkdir data/adult
mkdir data/compas
wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data -O data/adult/adult.data
wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.test -O data/adult/adult.test
wget https://raw.githubusercontent.com/propublica/compas-analysis/master/compas-scores.csv -O data/compas/compas-scores.csv

Train on Adult dataset with Alexr2

sh ./scripts/alexr2.sh

Train on Compas dataset with Alexr2

sh ./scripts/compas_alexr2.sh

Citation

If you find this tutorial helpful, please cite our paper:

@article{chen2025stochastic,
  title={Stochastic Momentum Methods for Non-smooth Non-Convex Finite-Sum Coupled Compositional Optimization},
  author={Chen, Xingyu and Wang, Bokun and Yang, Ming and Lin, Qihang and Yang, Tianbao},
  journal={arXiv preprint arXiv:2506.02504},
  year={2025}
}