Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep Learning
The autoencoder (AE) is a fundamental deep learning approach to anomaly detection. AEs are trained on the assumption that abnormal inputs will produce higher reconstruction errors than normal ones. In practice, however, this assumption is unreliable in the unsupervised case, where the training data...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9099561/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841554055990607872 |
|---|---|
| author | Nicholas Merrill Azim Eskandarian |
| author_facet | Nicholas Merrill Azim Eskandarian |
| author_sort | Nicholas Merrill |
| collection | DOAJ |
| description | The autoencoder (AE) is a fundamental deep learning approach to anomaly detection. AEs are trained on the assumption that abnormal inputs will produce higher reconstruction errors than normal ones. In practice, however, this assumption is unreliable in the unsupervised case, where the training data may contain anomalous examples. Given sufficient capacity and training time, an AE can generalize to such an extent that it reliably reconstructs anomalies. Consequently, the ability to distinguish anomalies via reconstruction errors is diminished. We respond to this limitation by introducing three new methods to more reliably train AEs for unsupervised anomaly detection: cumulative error scoring (CES), percentile loss (PL), and early stopping via knee detection. We demonstrate significant improvements over conventional AE training on image, remote-sensing, and cybersecurity datasets. |
| format | Article |
| id | doaj-art-a91a0da929bc4c28be80614ccd591673 |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-a91a0da929bc4c28be80614ccd5916732025-01-09T00:00:44ZengIEEEIEEE Access2169-35362020-01-01810182410183310.1109/ACCESS.2020.29973279099561Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep LearningNicholas Merrill0https://orcid.org/0000-0003-2217-4389Azim Eskandarian1Department of Mechanical Engineering, Virginia Tech, Blacksburg, VA, USADepartment of Mechanical Engineering, Virginia Tech, Blacksburg, VA, USAThe autoencoder (AE) is a fundamental deep learning approach to anomaly detection. AEs are trained on the assumption that abnormal inputs will produce higher reconstruction errors than normal ones. In practice, however, this assumption is unreliable in the unsupervised case, where the training data may contain anomalous examples. Given sufficient capacity and training time, an AE can generalize to such an extent that it reliably reconstructs anomalies. Consequently, the ability to distinguish anomalies via reconstruction errors is diminished. We respond to this limitation by introducing three new methods to more reliably train AEs for unsupervised anomaly detection: cumulative error scoring (CES), percentile loss (PL), and early stopping via knee detection. We demonstrate significant improvements over conventional AE training on image, remote-sensing, and cybersecurity datasets.https://ieeexplore.ieee.org/document/9099561/HyperspectralHSIdeep learninganomaly detectionunsupervisedautoencoder |
| spellingShingle | Nicholas Merrill Azim Eskandarian Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep Learning IEEE Access Hyperspectral HSI deep learning anomaly detection unsupervised autoencoder |
| title | Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep Learning |
| title_full | Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep Learning |
| title_fullStr | Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep Learning |
| title_full_unstemmed | Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep Learning |
| title_short | Modified Autoencoder Training and Scoring for Robust Unsupervised Anomaly Detection in Deep Learning |
| title_sort | modified autoencoder training and scoring for robust unsupervised anomaly detection in deep learning |
| topic | Hyperspectral HSI deep learning anomaly detection unsupervised autoencoder |
| url | https://ieeexplore.ieee.org/document/9099561/ |
| work_keys_str_mv | AT nicholasmerrill modifiedautoencodertrainingandscoringforrobustunsupervisedanomalydetectionindeeplearning AT azimeskandarian modifiedautoencodertrainingandscoringforrobustunsupervisedanomalydetectionindeeplearning |