HOSTILE DATA DETECTOR AND HOSTILE DATA DETECTION METHOD

To detect hostile data with high accuracy even under a situation where an attacker knows information about a method for detecting the hostile data, and thus maintain safety.SOLUTION: A hostile data detector holds evaluation target data and a model that outputs an evaluation value when data is input;...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: YOSHINO MASAYUKI, KAWANA NON, YAMAMOTO KYOHEI, YOKOHARI YUMIKO
Format: Patent
Sprache:eng ; jpn
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator YOSHINO MASAYUKI
KAWANA NON
YAMAMOTO KYOHEI
YOKOHARI YUMIKO
description To detect hostile data with high accuracy even under a situation where an attacker knows information about a method for detecting the hostile data, and thus maintain safety.SOLUTION: A hostile data detector holds evaluation target data and a model that outputs an evaluation value when data is input; inputs evaluation target data into the model to calculate an evaluation value of the evaluation target data; adds different noises to the evaluation target data to generate a plurality of extended evaluation target data; updates the extended evaluation target data by adding noise to each of the plurality of extended evaluation target data until the evaluation value of the extended evaluation target data does not match the evaluation value of the evaluation target data; and determines whether or not the evaluation target data is hostile data, based on the deviation of the evaluation values of the updated extended evaluation target data.SELECTED DRAWING: Figure 6 【課題】攻撃者が敵対的データの検知手法についての情報を知っている状況下においても高精度に敵対的データを検知して、ひいては安全性を保つ。【解決手段】敵対的データ検知装置は、評価対象データと、データが入力されると評価値を出力するモデルと、を保持し、評価対象データを前記モデルに入力して、評価対象データの評価値を算出し、評価対象データに異なるノイズを付加して、複数の拡張評価対象データを生成し、複数の拡張評価対象データそれぞれに対して、当該拡張評価対象データの評価値が、評価対象データの評価値と一致しないようになるまで、ノイズを加えて当該拡張評価対象データを更新し、更新後の拡張評価対象データそれぞれの評価値の偏りに基づいて、評価対象データが敵対的データであるかを判定する。【選択図】図6
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_JP2023056241A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>JP2023056241A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_JP2023056241A3</originalsourceid><addsrcrecordid>eNrjZDD38A8O8fRxVXBxDHFUcHENcXUO8Q9ScPRzUcAi4-nvp-DrGuLh78LDwJqWmFOcyguluRmU3FxDnD10Uwvy41OLCxKTU_NSS-K9AowMjIwNTM2MTAwdjYlSBABsAyhV</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>HOSTILE DATA DETECTOR AND HOSTILE DATA DETECTION METHOD</title><source>esp@cenet</source><creator>YOSHINO MASAYUKI ; KAWANA NON ; YAMAMOTO KYOHEI ; YOKOHARI YUMIKO</creator><creatorcontrib>YOSHINO MASAYUKI ; KAWANA NON ; YAMAMOTO KYOHEI ; YOKOHARI YUMIKO</creatorcontrib><description>To detect hostile data with high accuracy even under a situation where an attacker knows information about a method for detecting the hostile data, and thus maintain safety.SOLUTION: A hostile data detector holds evaluation target data and a model that outputs an evaluation value when data is input; inputs evaluation target data into the model to calculate an evaluation value of the evaluation target data; adds different noises to the evaluation target data to generate a plurality of extended evaluation target data; updates the extended evaluation target data by adding noise to each of the plurality of extended evaluation target data until the evaluation value of the extended evaluation target data does not match the evaluation value of the evaluation target data; and determines whether or not the evaluation target data is hostile data, based on the deviation of the evaluation values of the updated extended evaluation target data.SELECTED DRAWING: Figure 6 【課題】攻撃者が敵対的データの検知手法についての情報を知っている状況下においても高精度に敵対的データを検知して、ひいては安全性を保つ。【解決手段】敵対的データ検知装置は、評価対象データと、データが入力されると評価値を出力するモデルと、を保持し、評価対象データを前記モデルに入力して、評価対象データの評価値を算出し、評価対象データに異なるノイズを付加して、複数の拡張評価対象データを生成し、複数の拡張評価対象データそれぞれに対して、当該拡張評価対象データの評価値が、評価対象データの評価値と一致しないようになるまで、ノイズを加えて当該拡張評価対象データを更新し、更新後の拡張評価対象データそれぞれの評価値の偏りに基づいて、評価対象データが敵対的データであるかを判定する。【選択図】図6</description><language>eng ; jpn</language><subject>CALCULATING ; COMPUTING ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; PHYSICS</subject><creationdate>2023</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20230419&amp;DB=EPODOC&amp;CC=JP&amp;NR=2023056241A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25564,76547</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20230419&amp;DB=EPODOC&amp;CC=JP&amp;NR=2023056241A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>YOSHINO MASAYUKI</creatorcontrib><creatorcontrib>KAWANA NON</creatorcontrib><creatorcontrib>YAMAMOTO KYOHEI</creatorcontrib><creatorcontrib>YOKOHARI YUMIKO</creatorcontrib><title>HOSTILE DATA DETECTOR AND HOSTILE DATA DETECTION METHOD</title><description>To detect hostile data with high accuracy even under a situation where an attacker knows information about a method for detecting the hostile data, and thus maintain safety.SOLUTION: A hostile data detector holds evaluation target data and a model that outputs an evaluation value when data is input; inputs evaluation target data into the model to calculate an evaluation value of the evaluation target data; adds different noises to the evaluation target data to generate a plurality of extended evaluation target data; updates the extended evaluation target data by adding noise to each of the plurality of extended evaluation target data until the evaluation value of the extended evaluation target data does not match the evaluation value of the evaluation target data; and determines whether or not the evaluation target data is hostile data, based on the deviation of the evaluation values of the updated extended evaluation target data.SELECTED DRAWING: Figure 6 【課題】攻撃者が敵対的データの検知手法についての情報を知っている状況下においても高精度に敵対的データを検知して、ひいては安全性を保つ。【解決手段】敵対的データ検知装置は、評価対象データと、データが入力されると評価値を出力するモデルと、を保持し、評価対象データを前記モデルに入力して、評価対象データの評価値を算出し、評価対象データに異なるノイズを付加して、複数の拡張評価対象データを生成し、複数の拡張評価対象データそれぞれに対して、当該拡張評価対象データの評価値が、評価対象データの評価値と一致しないようになるまで、ノイズを加えて当該拡張評価対象データを更新し、更新後の拡張評価対象データそれぞれの評価値の偏りに基づいて、評価対象データが敵対的データであるかを判定する。【選択図】図6</description><subject>CALCULATING</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>ELECTRIC DIGITAL DATA PROCESSING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2023</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNrjZDD38A8O8fRxVXBxDHFUcHENcXUO8Q9ScPRzUcAi4-nvp-DrGuLh78LDwJqWmFOcyguluRmU3FxDnD10Uwvy41OLCxKTU_NSS-K9AowMjIwNTM2MTAwdjYlSBABsAyhV</recordid><startdate>20230419</startdate><enddate>20230419</enddate><creator>YOSHINO MASAYUKI</creator><creator>KAWANA NON</creator><creator>YAMAMOTO KYOHEI</creator><creator>YOKOHARI YUMIKO</creator><scope>EVB</scope></search><sort><creationdate>20230419</creationdate><title>HOSTILE DATA DETECTOR AND HOSTILE DATA DETECTION METHOD</title><author>YOSHINO MASAYUKI ; KAWANA NON ; YAMAMOTO KYOHEI ; YOKOHARI YUMIKO</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_JP2023056241A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>eng ; jpn</language><creationdate>2023</creationdate><topic>CALCULATING</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>YOSHINO MASAYUKI</creatorcontrib><creatorcontrib>KAWANA NON</creatorcontrib><creatorcontrib>YAMAMOTO KYOHEI</creatorcontrib><creatorcontrib>YOKOHARI YUMIKO</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>YOSHINO MASAYUKI</au><au>KAWANA NON</au><au>YAMAMOTO KYOHEI</au><au>YOKOHARI YUMIKO</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>HOSTILE DATA DETECTOR AND HOSTILE DATA DETECTION METHOD</title><date>2023-04-19</date><risdate>2023</risdate><abstract>To detect hostile data with high accuracy even under a situation where an attacker knows information about a method for detecting the hostile data, and thus maintain safety.SOLUTION: A hostile data detector holds evaluation target data and a model that outputs an evaluation value when data is input; inputs evaluation target data into the model to calculate an evaluation value of the evaluation target data; adds different noises to the evaluation target data to generate a plurality of extended evaluation target data; updates the extended evaluation target data by adding noise to each of the plurality of extended evaluation target data until the evaluation value of the extended evaluation target data does not match the evaluation value of the evaluation target data; and determines whether or not the evaluation target data is hostile data, based on the deviation of the evaluation values of the updated extended evaluation target data.SELECTED DRAWING: Figure 6 【課題】攻撃者が敵対的データの検知手法についての情報を知っている状況下においても高精度に敵対的データを検知して、ひいては安全性を保つ。【解決手段】敵対的データ検知装置は、評価対象データと、データが入力されると評価値を出力するモデルと、を保持し、評価対象データを前記モデルに入力して、評価対象データの評価値を算出し、評価対象データに異なるノイズを付加して、複数の拡張評価対象データを生成し、複数の拡張評価対象データそれぞれに対して、当該拡張評価対象データの評価値が、評価対象データの評価値と一致しないようになるまで、ノイズを加えて当該拡張評価対象データを更新し、更新後の拡張評価対象データそれぞれの評価値の偏りに基づいて、評価対象データが敵対的データであるかを判定する。【選択図】図6</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language eng ; jpn
recordid cdi_epo_espacenet_JP2023056241A
source esp@cenet
subjects CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
title HOSTILE DATA DETECTOR AND HOSTILE DATA DETECTION METHOD
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T08%3A11%3A16IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=YOSHINO%20MASAYUKI&rft.date=2023-04-19&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3EJP2023056241A%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true