Preset method

FlexMatch

Introduction

Based on the pseudo-labeling algorithm with fixed thresholds, FlexMatch proposes the course pseudo-labeling method (CPL) to estimate the learning effect of different categories in real time, and dynamically adjust the thresholds of different categories, so as to fully consider the learning difficulty between samples of different categories.

Citation

@article{zhang2021flexmatch,
  title={FlexMatch: Boosting Semi-supervised Learning with Curriculum Pseudo Labeling},
  author={Zhang, Bowen and Wang, Yidong and Hou, Wenxin and Wu, Hao and Wang, Jindong and Okumura, Manabu and Shinozaki, Takahiro},
  booktitle={Neural Information Processing Systems (NeurIPS)},
  year={2021}
}

SimMatch

Introduction

The instance similarity is introduced to record the features of labeled data, so that the unlabeled data can be compared with it, and the instance similarity predicted by the Teacher model can be corrected with the classification results, so as to obtain more accurate false labels.

Citation

@inproceedings{zheng2022simmatch,
  title={Simmatch: Semi-supervised learning with similarity matching},
  author={Zheng, Mingkai and You, Shan and Huang, Lang and Wang, Fei and Qian, Chen and Xu, Chang},
  booktitle={Computer Vision and Pattern Recognition},
  pages={14471--14481},
  year={2022}
}