/DCIC

Primary LanguagePython

DCIC

PMLR, 2020: Label-Noise Robust Domain Adaptation.

This is the basic implementation code (unofficial) for the paper: Label-Noise Robust Domain Adaptation.
Xiyu Yu, Tongliang Liu, Mingming Gong, Kun Zhang, Kayhan Batmanghelich, Dacheng Tao.

Dependencies

The methods are implemented by PyTorch on NVIDIA GeForce RTX 3080 Ti GPU. The environment is as bellow:

Install requirements.txt

pip install -r requirements.txt

Experiments

I verify the effectiveness of the proposed method on synthetic noisy datasets. In this repository, I provide the used datasets in the folder "data" (mnist.npz, usps.npz). Here is a training example:

python main.py \
    --source mnist \
    --target usps \
    --batch_size 128 \
    --noise_rate 0.2 \
    --random_state 1

If you find this code useful in your research, please cite

@inproceedings{liu2020dcic,
  title={Label-Noise Robust Domain Adaptation},
  author={Xiyu Yu, Tongliang Liu, Mingming Gong, Kun Zhang, Kayhan Batmanghelich, Dacheng Tao},
  booktitle={PMLR},
  year={2020}
}