L2E

An implementation for "A Unified Meta-Learning Framework for Dynamic Transfer Learning" (IJCAI'22) [Paper][arXiv].

Environment Requirements

The code has been tested under Python 3.6.5. The required packages are as follows:

• numpy==1.18.1
• torch==1.4.0
• torchvision==0.5.0
• higher==0.2.1
• Pillow==7.0.0

Data Sets

We used the following data sets in our experiments:

• Office-31
• Image-CLEF
• Caltran

Run the Codes

For dynamic transfer learning on a time evolving tasks on Office-31 or Image-CLEF, please run

python main.py

For dynamic transfer learning on a time evolving tasks on Caltran, please run

python utils/preprocess.py   ## pre-process Caltran to generate the time evolving tasks
python main_caltran.py

Acknowledgement

This is the latest source code of L2E for IJCAI-2022. If you find that it is helpful for your research, please consider to cite our paper:

@inproceedings{wu2022unified,
  title={A Unified Meta-Learning Framework for Dynamic Transfer Learning},
  author={Wu, Jun and He, Jingrui},
  booktitle={Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence},
  pages={3573--3579},
  year={2022}
}