/Mixup_for_UDA

Implementation of Adversarial Domain Adaptation with Domain Mixup (AAAI 2020 Oral).

Primary LanguagePython

Adversarial Domain Adaptation with Domain Mixup


This is the implementation of Adversarial Domain Adaptation with Domain Mixup in PyTorch. This work is accepted as Oral presentation at AAAI 2020.

Adversarial Domain Adaptation with Domain Mixup: [Paper (arxiv)].


Getting Started

  • We combine Domain Mixup strategy with a classical adversarial domain adaptation method, RevGrad, to showcase its effectiveness on boosting feature alignment. Details are presented in the Mixup_RevGrad folder.
  • The proposed DM-ADA approach utilizes a VAE-GAN based framework and performs Domain Mixup on both pixel and feature level. Details are presented in the DM-ADA folder. Some typical generations from source, target and mixup features are as follows (VisDA-2017 dataset is employed).

Citation

If this work helps your research, please cite the following paper.

@inproceedings{xu2020adversarial,
  author    = {Minghao Xu and Jian Zhang and Bingbing Ni and Teng Li and Chengjie Wang and Qi Tian and Wenjun Zhang},
  title     = {Adversarial Domain Adaptation with Domain Mixup},
  booktitle = {The Thirty-Fourth AAAI Conference on Artificial Intelligence},
  pages     = {6502--6509},
  publisher = {AAAI Press},
  year      = {2020}
}