Performance Enhancement of Autoencoder-Based Audio Data Anomaly Detection via Weakly Supervised Learning 


Vol. 48,  No. 3, pp. 382-390, Mar.  2023
10.7840/kics.2023.48.3.382


PDF
  Abstract

Recently, deep-learning based anomaly detection technology for a failure diagnosis of industrial facilities is receiving great attention. In general, the field of fault diagnosis and prediction has a characteristic that the number of abnormal data is significantly smaller than the number of normal data. Therefore, it is common to apply unsupervised learning assuming that most data are normal and train a model without labels. However, it is difficult to expect high accuracy in unsupervised learning owing to the lack of knowledge of anomaly features. In this paper paper, we propose a weakly supervised learning-based autoencoder that uses a small amount of abnormal data for learning features of anomalies, thus improve the performance of autoencoder. Experiments show that the proposed weakly supervised autoencoder is effective in improving the detection performance compared to the various conventional anomaly detection models while having low complexity.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Related Articles
  Cite this article

[IEEE Style]

Yong-GeunMoon, Min-SeongKwon, ByungjuLee, Jung-hoonNoh, "Performance Enhancement of Autoencoder-Based Audio Data Anomaly Detection via Weakly Supervised Learning," The Journal of Korean Institute of Communications and Information Sciences, vol. 48, no. 3, pp. 382-390, 2023. DOI: 10.7840/kics.2023.48.3.382.

[ACM Style]

Yong-GeunMoon, Min-SeongKwon, ByungjuLee, and Jung-hoonNoh. 2023. Performance Enhancement of Autoencoder-Based Audio Data Anomaly Detection via Weakly Supervised Learning. The Journal of Korean Institute of Communications and Information Sciences, 48, 3, (2023), 382-390. DOI: 10.7840/kics.2023.48.3.382.

[KICS Style]

Yong-GeunMoon, Min-SeongKwon, ByungjuLee, Jung-hoonNoh, "Performance Enhancement of Autoencoder-Based Audio Data Anomaly Detection via Weakly Supervised Learning," The Journal of Korean Institute of Communications and Information Sciences, vol. 48, no. 3, pp. 382-390, 3. 2023. (https://doi.org/10.7840/kics.2023.48.3.382)
Vol. 48, No. 3 Index

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880