Patch-Based Time-Series Anomaly Detection with Cross-Variable Attention

Jeena Son, Seunghwan Song, Jun Geol Baek

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    Modern manufacturing processes are influenced by smart factories, which collect data from multiple sensors in real time. However, detecting anomalies is challenging due to their irregular and complex patterns, often resulting in unreliable labeling that depends on the engineer's expertise. Moreover, in these processes, where sensors are interconnected, fluctuations in one variable can potentially influence subsequent variables. The proposed method uses cross-variable attention to reflect the relationship between variables at different time points. It also utilizes contrastive learning to extract a representation of only normal data that can be distinguished from anomalies without labels. Experimental results on multivariate time series data demonstrate that the proposed method outperforms existing models in anomaly detection.

    Original languageEnglish
    Title of host publication6th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2024
    PublisherInstitute of Electrical and Electronics Engineers Inc.
    Pages638-643
    Number of pages6
    ISBN (Electronic)9798350344349
    DOIs
    Publication statusPublished - 2024
    Event6th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2024 - Osaka, Japan
    Duration: 2024 Feb 192024 Feb 22

    Publication series

    Name6th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2024

    Conference

    Conference6th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2024
    Country/TerritoryJapan
    CityOsaka
    Period24/2/1924/2/22

    Bibliographical note

    Publisher Copyright:
    © 2024 IEEE.

    Keywords

    • Anomaly Detection
    • Contrastive Learning
    • Cross-Variable Attention
    • Multivariate Time Series

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Computer Networks and Communications
    • Computer Science Applications
    • Computer Vision and Pattern Recognition
    • Information Systems
    • Safety, Risk, Reliability and Quality
    • Health Informatics

    Fingerprint

    Dive into the research topics of 'Patch-Based Time-Series Anomaly Detection with Cross-Variable Attention'. Together they form a unique fingerprint.

    Cite this