LEARNING DEVICE, CLASSIFICATION DEVICE, LEARNING METHOD, CLASSIFICATION METHOD, AND PROGRAM

A learning device according to an embodiment is characterized by comprising: an input means which enters both training data for learning a classifier, and a cause-and-effect graph representing the cause-and-effect relationship between variables included in the training data; and a learning means whi...

Full description

Saved in:
Bibliographic Details
Main Authors CHIKAHARA, Yoichi, FUJINO, Akinori
Format Patent
LanguageEnglish
French
Japanese
Published 06.05.2021
Subjects
Online AccessGet full text

Cover

Loading…
Abstract A learning device according to an embodiment is characterized by comprising: an input means which enters both training data for learning a classifier, and a cause-and-effect graph representing the cause-and-effect relationship between variables included in the training data; and a learning means which uses the training data and the cause-and-effect graph entered by the input means to learn the classifier by solving an optimization problem that includes the constraint that the average of the cause-and-effect impacts between prescribed variables be within a prescribed range and the variance of the cause-and-effect impacts be equal to or less than a prescribed value. Selon un mode de réalisation, la présente invention concerne un dispositif d'apprentissage caractérisé en ce qu'il comprend : un moyen d'entrée qui saisit à la fois des données d'apprentissage permettant d'apprendre un classifieur, et un graphique cause-effet représentant la relation de cause à effet entre des variables étant dans les données d'apprentissage ; et un moyen d'apprentissage qui utilise les données d'apprentissage et le graphique de cause-effet saisi par le moyen d'entrée afin d'apprendre le classifieur en résolvant un problème d'optimisation qui comporte la contrainte que la moyenne des impacts de cause à effet entre des variables prescrites se trouve dans une plage prescrite et que la variance des impacts de cause à effet soit égale ou inférieure à une valeur prescrite. 一実施形態に係る学習装置は、分類器を学習するための訓練データと、前記訓練データに含まれる変数間の因果関係を表す因果グラフとを入力する入力手段と、前記入力手段により入力された訓練データと因果グラフとを用いて、所定の変数間の因果効果の平均が所定の範囲内にあり、かつ、前記因果効果の分散が所定の値以下である制約付き最適化問題を解くことにより前記分類器を学習する学習手段と、を有することを特徴とする。
AbstractList A learning device according to an embodiment is characterized by comprising: an input means which enters both training data for learning a classifier, and a cause-and-effect graph representing the cause-and-effect relationship between variables included in the training data; and a learning means which uses the training data and the cause-and-effect graph entered by the input means to learn the classifier by solving an optimization problem that includes the constraint that the average of the cause-and-effect impacts between prescribed variables be within a prescribed range and the variance of the cause-and-effect impacts be equal to or less than a prescribed value. Selon un mode de réalisation, la présente invention concerne un dispositif d'apprentissage caractérisé en ce qu'il comprend : un moyen d'entrée qui saisit à la fois des données d'apprentissage permettant d'apprendre un classifieur, et un graphique cause-effet représentant la relation de cause à effet entre des variables étant dans les données d'apprentissage ; et un moyen d'apprentissage qui utilise les données d'apprentissage et le graphique de cause-effet saisi par le moyen d'entrée afin d'apprendre le classifieur en résolvant un problème d'optimisation qui comporte la contrainte que la moyenne des impacts de cause à effet entre des variables prescrites se trouve dans une plage prescrite et que la variance des impacts de cause à effet soit égale ou inférieure à une valeur prescrite. 一実施形態に係る学習装置は、分類器を学習するための訓練データと、前記訓練データに含まれる変数間の因果関係を表す因果グラフとを入力する入力手段と、前記入力手段により入力された訓練データと因果グラフとを用いて、所定の変数間の因果効果の平均が所定の範囲内にあり、かつ、前記因果効果の分散が所定の値以下である制約付き最適化問題を解くことにより前記分類器を学習する学習手段と、を有することを特徴とする。
Author CHIKAHARA, Yoichi
FUJINO, Akinori
Author_xml – fullname: CHIKAHARA, Yoichi
– fullname: FUJINO, Akinori
BookMark eNrjYmDJy89L5WSI9nF1DPLz9HNXcHEN83R21VFw9nEMDvZ083R2DPH094MLw9X5uoZ4-LtgqIMJO_q5KAQE-bsHOfryMLCmJeYUp_JCaW4GZTfXEGcP3dSC_PjU4oLE5NS81JL4cH8jAyNDAwsTMwNLR0Nj4lQBAFfANEU
ContentType Patent
DBID EVB
DatabaseName esp@cenet
DatabaseTitleList
Database_xml – sequence: 1
  dbid: EVB
  name: esp@cenet
  url: http://worldwide.espacenet.com/singleLineSearch?locale=en_EP
  sourceTypes: Open Access Repository
DeliveryMethod fulltext_linktorsrc
Discipline Medicine
Chemistry
Sciences
Physics
DocumentTitleAlternate 学習装置、分類装置、学習方法、分類方法及びプログラム
DISPOSITIF D'APPRENTISSAGE, DISPOSITIF DE CLASSIFICATION, PROCÉDÉ D'APPRENTISSAGE, PROCÉDÉ DE CLASSIFICATION ET PROGRAMME
ExternalDocumentID WO2021084609A1
GroupedDBID EVB
ID FETCH-epo_espacenet_WO2021084609A13
IEDL.DBID EVB
IngestDate Fri Jul 19 14:33:45 EDT 2024
IsOpenAccess true
IsPeerReviewed false
IsScholarly false
Language English
French
Japanese
LinkModel DirectLink
MergedId FETCHMERGED-epo_espacenet_WO2021084609A13
Notes Application Number: WO2019JP42339
OpenAccessLink https://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20210506&DB=EPODOC&CC=WO&NR=2021084609A1
ParticipantIDs epo_espacenet_WO2021084609A1
PublicationCentury 2000
PublicationDate 20210506
PublicationDateYYYYMMDD 2021-05-06
PublicationDate_xml – month: 05
  year: 2021
  text: 20210506
  day: 06
PublicationDecade 2020
PublicationYear 2021
RelatedCompanies NIPPON TELEGRAPH AND TELEPHONE CORPORATION
RelatedCompanies_xml – name: NIPPON TELEGRAPH AND TELEPHONE CORPORATION
Score 3.455865
Snippet A learning device according to an embodiment is characterized by comprising: an input means which enters both training data for learning a classifier, and a...
SourceID epo
SourceType Open Access Repository
SubjectTerms CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
PHYSICS
Title LEARNING DEVICE, CLASSIFICATION DEVICE, LEARNING METHOD, CLASSIFICATION METHOD, AND PROGRAM
URI https://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20210506&DB=EPODOC&locale=&CC=WO&NR=2021084609A1
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV3rS8MwED_GfH7TqfiYUlD6yeLsa_bDkC5Jt8n6oMw58MPoExTZhq3473uNbR0I-5hfjpAc_HK55O4CcJPch900TlXJkNMHSY2DFCklK5KmRKESpXqYJMXVgO3ow2f1aabNGvBR5cLwOqHfvDgiMipCvud8v179XWJRHluZ3YVvCC0frUmPiqV3jP6L1tFF2u8xz6UuEQlBv010_N8-tLUdw0RfaQsP0t2CD2zaL_JSVutGxTqAbQ_HW-SH0HgPWrBHqr_XWrBrl0_eLdjhMZpRhmDJw-wIXsfM9J2RMxAom44IuxXI2MR90Srzgmu4lrPZZOjSf3IVbDpU8Hx34Jv2MVxbbEKGEs53Xqtn_uKuL045geZiuUhOQUh0oyi1YvCqWkasBYYSFPZbV0I8hkXyGbQ3jXS-ufsC9osmD_7T29DMP7-SSzTQeXjF9foDHYWKzQ
link.rule.ids 230,309,783,888,25576,76876
linkProvider European Patent Office
linkToHtml http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV3dS8NADA9jfsw3nYofUwtKnyzOde3sw5Du7rpO-zFKnYM9jLZrQZFtuIr_vunZ1oGw1yQcd4FfcsklOYCb-D7sJLOkLWmt5EFqz4IEIdWSJUWOQjlK1DCOs9SA7ajmS_tprIwr8FH0wvA5od98OCIiKkK8p9xeL_-SWJTXVq7uwjckLR4Nv0vFPDrG-EVpqiLtddnQpS4RCcG4TXS8Xx762qamY6y0hZfsToYHNuplfSnLdadi7MP2ENebpwdQeQ_qUCPF32t12LXzJ-867PAazWiFxByHq0OYWEz3nIHTFygbDQi7FYilo1008r7gklzK2cw3XfpPriDrDhWGntv3dPsIrg3mE1PC_U5L9Uxf3fXDycdQnS_m8QkIsaplo1Y0PlVLmymBJgeZ_1blEK9hUesUGptWOtvMvoKa6dvW1Bo4z-ewl7F4IaDagGr6-RVfoLNOw0uu4x_NIY3A
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Apatent&rft.title=LEARNING+DEVICE%2C+CLASSIFICATION+DEVICE%2C+LEARNING+METHOD%2C+CLASSIFICATION+METHOD%2C+AND+PROGRAM&rft.inventor=CHIKAHARA%2C+Yoichi&rft.inventor=FUJINO%2C+Akinori&rft.date=2021-05-06&rft.externalDBID=A1&rft.externalDocID=WO2021084609A1