ALBA: Active Learning Bandits

This repository contains the code and resources accompanying the Multi-domain Active Learning for Semi-supervised Anomaly Detection research paper, accepted for publication at the ECML PKDD 2022 conference in Grenoble.

@inproceedings{vercruyssen2022multi,
  title={Multi-domain Active Learning for Semi-supervised Anomaly Detection},
  author={Vercruyssen, Vincent and Perini, Lorenzo and Meert, Wannes and Davis, Jesse},
  booktitle={Joint European Conference on Machine Learning and Knowledge Discovery in Databases},
  year={2022},
  organization={Springer}
}

Resources

You can find the paper, the slides presented at ECML PKDD, and the poster presented at ECML PKDD in this repository.

Code

[TO BE UPDATED] In the src/ folder, you will find a minimal implementation of:

1. The ALBA active learning algorithm (our main contribution).
2. The baselines compared in the paper.
3. Instructions on how to use the algorithms and an illustration on toy data.

Images

In the images/ folder, you will find pdf files for some of the main images showed in the paper.

Data - disclaimer

The datasets used in the experiments of the paper are proprietary and can unfortunately not be shared at the moment.

Contact and support

For any question, you can contact Vincent Vercruyssen, or any of the other authors.