Deep Transfer Learning in PyTorch

This is a PyTorch library for deep transfer learning. We divide the code into two aspects: Single-source Unsupervised Domain Adaptation (SUDA) and Multi-source Unsupervised Domain Adaptation (MUDA). There are many SUDA methods, however I find there is a few MUDA methods with deep learning. Besides, MUDA with deep learning might be a more promising direction for domain adaptation.

Here I have implemented some deep transfer methods as follows:

• UDA
  • DDC：Deep Domain Confusion Maximizing for Domain Invariance
  • DAN: Learning Transferable Features with Deep Adaptation Networks (ICML2015)
  • Deep Coral: Deep CORAL Correlation Alignment for Deep Domain Adaptation (ECCV2016)
  • Revgrad: Unsupervised Domain Adaptation by Backpropagation (ICML2015)
  • MRAN: Multi-representation adaptation network for cross-domain image classification (Neural Network 2019)
  • DSAN: Deep Subdomain Adaptation Network for Image Classification (IEEE Transactions on Neural Networks and Learning Systems 2020)
• MUDA
  • Aligning Domain-specific Distribution and Classifier for Cross-domain Classification from Multiple Sources (AAAI2019)
• Application
  • Cross-domain Fraud Detection: Modeling Users’ Behavior Sequences with Hierarchical Explainable Network for Cross-domain Fraud Detection (WWW2020)
  • Learning to Expand Audience via Meta Hybrid Experts and Critics for Recommendation and Advertising (KDD2021)
• Survey
  • A Comprehensive Survey on Transfer Learning (Proc. IEEE)

Results on Office31(UDA)

Note that the results without '*' comes from paper. The results with '*' are run by myself with the code.

Results on Office31(MUDA)

Results on OfficeHome(MUDA)

Note that (1) Source combine: all source domains are combined together into a traditional single-source v.s. target setting. (2) Single best: among the multiple source domains, we report the best single source transfer results. (3) Multi-source: the results of MUDA methods.

Note

If you find that your accuracy is 100%, the problem might be the dataset folder. Please note that the folder structure required for the data provider to work is:

-dataset
    -amazon
    -webcam
    -dslr

Contact

If you have any problem about this library, please create an Issue or send us an Email at:

• [email protected]
• [email protected]

Reference

If you use this repository, please cite the following papers:

@inproceedings{zhu2019aligning,
  title={Aligning domain-specific distribution and classifier for cross-domain classification from multiple sources},
  author={Zhu, Yongchun and Zhuang, Fuzhen and Wang, Deqing},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={33},
  pages={5989--5996},
  year={2019}
}

@article{zhu2020deep,
  title={Deep subdomain adaptation network for image classification},
  author={Zhu, Yongchun and Zhuang, Fuzhen and Wang, Jindong and Ke, Guolin and Chen, Jingwu and Bian, Jiang and Xiong, Hui and He, Qing},
  journal={IEEE transactions on neural networks and learning systems},
  volume={32},
  number={4},
  pages={1713--1722},
  year={2020},
  publisher={IEEE}
}

@article{zhu2019multi,
  title={Multi-representation adaptation network for cross-domain image classification},
  author={Zhu, Yongchun and Zhuang, Fuzhen and Wang, Jindong and Chen, Jingwu and Shi, Zhiping and Wu, Wenjuan and He, Qing},
  journal={Neural Networks},
  volume={119},
  pages={214--221},
  year={2019},
  publisher={Elsevier}
}