Title:
LEARNING DEVICE, LEARNING METHOD, COMPUTER PROGRAM, AND RECORDING MEDIUM
Document Type and Number:
WIPO Patent Application WO/2020/189132
Kind Code:
A1
Abstract:
This learning device comprises: a generation means for generating a learning model that performs machine learning using normal data that is time-series data indicating a normal state and thereby predicts and outputs time-series data that corresponds to inputted time-series data; a first acquisition means for comparing predicted normal data for a second period that is predicted by inputting, to the learning model, normal data for a first period with normal data for the second period, and acquiring a first deviance; a second acquisition means for comparing predicted abnormal data for a fourth period that is predicted by inputting, to the learning model, abnormal data that is time-series data indicating an abnormal state in a third period with abnormal data for the fourth period, and acquiring a second deviance; and a detection means for detecting, on the basis of the first deviance and the second deviance, a time-series pattern that is lacking among time-series patterns indicating a normal state pertaining to the normal data.
Inventors:
IIZAWA YOHEI (JP)
Application Number:
PCT/JP2020/006039
Publication Date:
September 24, 2020
Filing Date:
February 17, 2020
Export Citation:
Assignee:
NEC CORP (JP)
International Classes:
G06N20/00
Other References:
TSUKADA, MINETO ET AL.: "A Stable and Efficient Learning Method for FPGA-Based Online Sequential Unsupervised Anomaly Detector", IEICE TECHNICAL REPORT, vol. 118, no. 165, 23 July 2018 (2018-07-23), pages 217 - 222, ISSN: 2432-6380
MALHOTRA, PANKAJ ET AL.: "Long Short Term Memory Networks for Anomaly Detection in Time Series", PROCEEDINGS OF THE 23RD EUROPEAN SYMPOSIUM ON ARTIFICIAL NEURAL NETWORKS, 2015, pages 89 - 94, XP055429593, Retrieved from the Internet [retrieved on 20200331]
MALHOTRA, PANKAJ ET AL.: "Long Short Term Memory Networks for Anomaly Detection in Time Series", PROCEEDINGS OF THE 23RD EUROPEAN SYMPOSIUM ON ARTIFICIAL NEURAL NETWORKS, 2015, pages 89 - 94, XP055429593, Retrieved from the Internet
Attorney, Agent or Firm:
EGAMI, TATSUO et al. (JP)
Download PDF:
Previous Patent: DISPLAY DEVICE
Next Patent: SYSTEM, CLIENT DEVICE, DATA PROCESSING METHOD, AND COMPUTER PROGRAM
Next Patent: SYSTEM, CLIENT DEVICE, DATA PROCESSING METHOD, AND COMPUTER PROGRAM