Experiments for Adversarial Invariant Feature Learning under Accuracy Constraint for Domain Generalization

Paper Info

• arxiv
• supplementary: ./docs/supplementary.pdf

Requirements

• python 2.7
• numpy==1.12.1
• pandas==0.20.1
• sacred==0.7.4
• tensorboardX==1.2
• torch==0.3.1

Implementation

• $ python dataset.py
  • The MNISTR dataset will be automatically downloaded under ~/.torch/datasets/ by running the script.
• $ ./run.sh
• Then you can make a score table in ./analyze_results.ipynb

References

• Ghifary, M., Bastiaan Kleijn, W., Zhang, M., Balduzzi, D.: Domain generalization for object recognition with multi-task autoencoders. In: Proc. of the IEEE International Conference on Computer Vision (ICCV) (2015)
  • We borrow the dataset from their repository https://github.com/ghif/mtae