Robust Positive-Unlabeled Learning via Self-selection of Unlabeled Negatives

This repo contains the official PyTorch code for Robust-PU.

Introduction

We utilize a novel "hardness" measure to distinguish unlabeled samples with a high chance of being negative from unlabeled samples with large label noise. An iterative training strategy is then implemented to fine-tune the selection of negative samples during the training process in an iterative manner to include more "easy" samples in the early stage of training.

Usage

Requirements

Python 3.9

conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch

```
pip install -r requirements.txt
```

Experiments

To train the model on MNIST with prior 0.6:

python main.py --dataset mnist --prior 0.6 --pre_lr 1e-3 --pre_epochs 100 --pre_batch_size 128 --pre_wd 1e-4 --lr 1e-2 --inner_epochs 20 --wd 0 --batch_size 64 --scheduler_type_n linear --alpha_n 0.11 --max_thresh_n 1.0 --grow_steps_n 5 --temper_n 1.3 --scheduler_type_p linear --alpha_p 0.1 --max_thresh_p 1.0 --grow_steps_p 5 --temper_p 1.0 --hardness logistic

${scheduler_type_n}: const, linear, convex, concave, exp

${scheduler_type_p}: const, linear, convex, concave, exp

${hardness}: logistic, sigmoid

Contact

If you have any question, please feel free to contact the authors. Zhangchi Zhu: zczhu@stu.ecnu.edu.cn

Citation

If you find our work is useful in your research, please consider citing: