Home

Awesome

Neural Transformation Learning for Anomaly Detection (NeuTraLAD)

This is the companion code for a PyTorch implementation of Neural Transformation Learning reported in the paper Neural Transformation Learning for Deep Anomaly Detection Beyond Images by Chen Qiu et al. The paper is published in ICML 2021 and can be found here https://arxiv.org/abs/2103.16440. The code allows the users to reproduce and extend the results reported in the study. Please cite the above paper when reporting, reproducing or extending the results.

Purpose of the project

This software is a research prototype, solely developed for and published as part of the publication cited above. It will neither be maintained nor monitored in any way.

Reproduce the Results

This repo contains the code of experiments with NeuTraLAD on various data types including time series data (one-vs-rest setting), tabular data, image data (one-vs-rest setting), text data (one-vs-rest setting), and graph data (one-vs-rest setting).

Please run the command and replace $# with available options (see below):

python Launch_Exps.py --config-file $1 --dataset-name $2 

config-file:

dataset-name:

How to Use

  1. When using your own data, please put your data files under DATA.

  2. Create a config file which contains your hyper-parameters under config_files.

  3. Add your data loader to the function ''load_data'' in the loader/LoadData.py.

Datasets

License

Neural Transformation Learning for Anomaly Detection (NeuTraLAD) is open-sourced under the AGPL-3.0 license. See the LICENSE file for details.

For a list of other open source components included in Neural Transformation Learning for Anomaly Detection (NeuTraLAD), see the file 3rd-party-licenses.txt.