We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Anomaly Detection Based on Multiple-Hypothesis Autoencoder

Abstract: Recently Autoencoder(AE) based models are widely used in the field of anomaly detection. A model trained with normal data generates a larger restoration error for abnormal data. Whether or not abnormal data is determined by observing the restoration error. It takes a lot of cost and time to obtain abnormal data in the industrial field. Therefore the model trains only normal data and detects abnormal data in the inference phase. However, the restoration area for the input data of AE is limited in the latent space. To solve this problem, we propose Multiple-hypothesis Autoencoder(MH-AE) model composed of several decoders. MH-AE model increases the restoration area through contention between decoders. The proposed method shows that the anomaly detection performance is improved compared to the traditional AE for various input datasets.
Comments: 3pages, 3figures
Subjects: Machine Learning (cs.LG); Signal Processing (eess.SP)
Cite as: arXiv:2107.08790 [cs.LG]
  (or arXiv:2107.08790v1 [cs.LG] for this version)

Submission history

From: JoonSung Lee [view email]
[v1] Wed, 7 Jul 2021 05:09:03 GMT (306kb,D)

Link back to: arXiv, form interface, contact.