Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
LEARNING DEVICE, ABNORMALITY DETECTION DEVICE, LEARNING METHOD, AND ABNORMALITY DETECTION METHOD
Document Type and Number:
WIPO Patent Application WO/2021/214833
Kind Code:
A1
Abstract:
A learning device according to one embodiment is characterizing by having: an input unit for inputting a dataset of data that represents an aggregate composed of pairs of an attribute value and attribute information; a computation unit for calculating, using an aggregate of vectors which are a combination of the attribute value with an embedded vector in which the attribute information is embedded, a prediction value that corresponds to the attribute value of each piece of data included in the dataset; and a learning unit for updating parameters that include the embedded vectors so as to minimize a difference between the attribute value and the prediction value corresponding to the attribute value.

Inventors:
TAJIRI KENGO (JP)
IWATA TOMOHARU (JP)
WATANABE KEISHIRO (JP)
Application Number:
PCT/JP2020/017062
Publication Date:
October 28, 2021
Filing Date:
April 20, 2020
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
NIPPON TELEGRAPH & TELEPHONE (JP)
International Classes:
G06N20/00
Foreign References:
US20180203921A12018-07-19
US20180204120A12018-07-19
US20190132343A12019-05-02
Other References:
MISHRA ASHISH, REDDY SHIVA KRISHNA, MITTAL ANURAG, MURTHY HEMA A.: "A Generative Model for Zero Shot Learning Using Conditional Variational Autoencoders", ARXIV.ORG, IEEE, 27 January 2018 (2018-01-27), pages 1 - 9, XP055867430, ISBN: 978-1-5386-6100-0, Retrieved from the Internet [retrieved on 20211130], DOI: 10.1109/CVPRW.2018.00294
LEE JUHO, LEE YOONHO, KIM JUNGTAEK, KOSIOREK ADAM R, CHOI SEUNGJIN, TEH YEE WHYE: "Set Transformer: A Framework for Attention-based Permutation-Invariant Neural Networks", ARXIV.ORG, 26 May 2019 (2019-05-26), pages 1 - 17, XP055867435, Retrieved from the Internet [retrieved on 20211130]
Attorney, Agent or Firm:
ITOH, Tadashige et al. (JP)
Download PDF: