https://github.com/borealisai/robust-gan
On Minimax Optimality of GANs for Robust Mean Estimation
https://github.com/borealisai/robust-gan
Last synced: 5 months ago
JSON representation
On Minimax Optimality of GANs for Robust Mean Estimation
- Host: GitHub
- URL: https://github.com/borealisai/robust-gan
- Owner: BorealisAI
- License: other
- Created: 2020-04-03T18:50:44.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2024-07-25T10:56:42.000Z (almost 2 years ago)
- Last Synced: 2025-01-22T03:33:05.191Z (over 1 year ago)
- Language: Python
- Homepage:
- Size: 20.5 KB
- Stars: 1
- Watchers: 7
- Forks: 1
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# What is this repository?
This repository contains code for paper "[On Minimax Optimality of GANs for Robust Mean Estimation][paper]" (AISTATS 2020).
[paper]: https://cs.uwaterloo.ca/~k77wu/paper/aistats2020.pdf
We implemented f-GAN, MMD-GAN (with Gaussian kernel) and Wasserstein GAN (with Euclidean norm as ground cost). These models are tested under Huber's contamination model.
## Usage
To install dependency, run
```
pip install -r requirements.txt
```
Run the following scripts containing detailed parameter configurations:
```
bash test_fgan.sh
bash test_mmd.sh
bash test_sinkhorn.sh
```