Anomaly Detection from a Tensor Train Perspective
This addresses anomaly detection problems for domains like cybersecurity and image analysis, but appears incremental as it applies existing tensor network methods to new datasets.
The paper tackled anomaly detection by using tensor train representations to compress normal data and delete anomalous data structure, achieving effective results on digits, Olivetti faces, and cybersecurity datasets.
We present a series of algorithms in tensor networks for anomaly detection in datasets, by using data compression in a Tensor Train representation. These algorithms consist of preserving the structure of normal data in compression and deleting the structure of anomalous data. The algorithms can be applied to any tensor network representation. We test the effectiveness of the methods with digits and Olivetti faces datasets and a cybersecurity dataset to determine cyber-attacks.