Semi-Supervised Variational Adversarial Active Learning via Learning to Rank and Agreement-Based Pseudo Labeling
Active learning aims to alleviate the amount of labor involved in data labeling by automating the selection of unlabeled samples via an acquisition function. For example, variational adversarial active learning (VAAL) leverages an adversarial network to discriminate unlabeled samples from labeled ones using latent space information. However, VAAL has the following shortcomings: (i) it does not exploit target task information, and (ii) unlabeled data is only used for sample selection rather than model training. To address these limitations, we introduce novel techniques that significantly improve the use of abundant unlabeled data during training and take into account the task information.
This repository provides source code for our 2024 ICPR paper titled "Semi-Supervised Variational Adversarial Active Learning via Learning to Rank and Agreement-Based Pseudo Labeling." Concretely, we propose an improved pseudo-labeling algorithm that leverages information from all unlabeled data in a semi-supervised manner, thus allowing a model to explore a richer data space. In addition, we develop a ranking-based loss prediction module that converts predicted relative ranking information into a differentiable ranking loss. This loss can be embedded as a rank variable into the latent space of a variational autoencoder and then trained with a discriminator in an adversarial fashion for sample selection. We demonstrate the superior performance of our approach over the state of the art on various image classification and segmentation benchmark datasets.
If you find this project useful, then please consider citing our work.
@inproceedings{lyu2024semi,
title={Semi-Supervised Variational Adversarial Active Learning via Learning to Rank and Agreement-Based Pseudo Labeling},
author={Lyu, Zongyao and Beksi, William J},
booktitle={Proceedings of the International Conference on Pattern Recognition (ICPR)},
pages={},
year={2024}
}
Run the following command to install the required packages.
pip install -r requirements.txt We utilize the SoDeep software to provide a ranking algorithm as the sorter in our framework. This sorter is trained separately from the rest of the model. To train the sorter, you can go to the SoDeep repository and follow the instructions. In this work, we use the LSTM sorter with a length of 128. After training is complete, the sorter is applied to the loss prediction module to convert the predicted and target losses into ranking losses for the active learning process. (Note: You can train a sorter with any length, but the length of the sorter must be consistent with the batch size.)
Run the following script to train the active learning model.
python main_sodeep_semi.py --dataset cifar10 --data_path path/to/your/data --batch_size 128 --no_of_epochs 150 --trials 1 --cycles 10 --weight_path path/to/your/sorter
