By Risheng Liu^1*,2, Jiaxin Gao¹, Xuan Liu¹, Xin Fan¹

¹Dalian University of Technology, ²Peng Cheng Laboratory

You can simply run the following command automatically install the dependencies

pip install -r requirement.txt

This code mainly requires the following:

• Python 3.*
• tqdm
• Pytorch
• higher

You can run the python file for different applications following the script below:

1. GAN and Its Variants

Python unrolled_gan_rhg_ring_LwCL.py  # For 2D Ring MOG dataset.
Python  unrolled_gan_rhg_cube_LwCL.py  # For 3D Cube MOG dataset.

2. Multi-Task Meta-Learning Few-Shot Classification

For the few-shot classification experiments in multi-task meta-learning, the entire network architecture is based on the L2F network. You can download the complete code from Baidu Yun (extraction code: i06p). The datasets used are mini-Imagenet and Omniglot. Please download the corresponding dataset, for example, mini_imagenet_full_size.tar.bz2, and place it in the dataset directory. Then, execute the following command:

python train_maml_system.py --name_of_args_json_file experiment_config/mini-imagenet_l2f_mini-imagenet_5_way_1_shot_0_resnet12_GN.json --gpu_to_use 0  # For few-shot classification tasks.

3. Hyper-Parameter Learning

For hyper-cleaning experiments, download the corresponding datasets, and place it in the dataset directory. Then, execute the following command:

python ./HPL/data_hyper_cleaning.py  # For data hyper-cleaning tasks.

1. Numerical mechanism evaluation:

If you feel this project is helpful, please consider cite our paper 😊

@article{liu2024learning,
  title={Learning with Constraint Learning: New Perspective, Solution Strategy and Various Applications},
  author={Liu, Risheng and Gao, Jiaxin and Liu, Xuan and Fan, Xin},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  volume={46},
  number={7},
  pages={5026--5043},
  year={2024},
  publisher={IEEE}

}

@article{liu2021investigating,
  title={Investigating bi-level optimization for learning and vision from a unified perspective: A survey and beyond},
  author={Liu, Risheng and Gao, Jiaxin and Zhang, Jin and Meng, Deyu and Lin, Zhouchen},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  volume={44},
  number={12},
  pages={10045--10067},
  year={2021},
  publisher={IEEE}
}

Part of the code is adapted from previous works: IAPTT-GM, L2F and BLO. We thank the authors for sharing the codes for their great works.

If you have any inquiries, feel free to reach out to Jiaxin Gao via email at jiaxinn.gao@outlook.com, or contact Xuan Liu at liuxuan_16@126.com.