Model security detection method based on generative adversarial network

The invention discloses a model security detection method based on a generative adversarial network. The method comprises the following specific steps of: 1, setting a behavior similarity security threshold value delta; 2, constructing an initialization generator G and a substitution model D; 3, car...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: MIAO HONGLE, GAO YING, WU HONGRUI, CHEN JIXIANG
Format: Patent
Sprache:chi ; eng
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 MIAO HONGLE
GAO YING
WU HONGRUI
CHEN JIXIANG
description The invention discloses a model security detection method based on a generative adversarial network. The method comprises the following specific steps of: 1, setting a behavior similarity security threshold value delta; 2, constructing an initialization generator G and a substitution model D; 3, carrying out iteration round by round, wherein the behavior similarity mu between the substitution model D and the tested model T is calculated, and the next step is carried out after the behavior similarity mu reaches a set value; and 4, evaluating the safety of the tested model T. The model stealing method based on the GAN is suitable for a black box attack scene without training data, and the behavior similarity of the substitution model and the tested model is rapidly improved by generating artificial data with relatively balanced categories. According to experimental results, the method has the characteristics of high adaptability, high efficiency and the like, and the risk of model stealing attack after the acce
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_CN112989361A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>CN112989361A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_CN112989361A3</originalsourceid><addsrcrecordid>eNrjZHD3zU9JzVEoTk0uLcosqVRISS1JTS7JzM9TyE0tychPUUhKLE5NUQDy01PzUosSSzLLUhUSU8pSi4oTizITcxTyUkvK84uyeRhY0xJzilN5oTQ3g6Kba4izh25qQX58anFBYjJQe0m8s5-hoZGlhaWxmaGjMTFqAKRoNIU</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Model security detection method based on generative adversarial network</title><source>esp@cenet</source><creator>MIAO HONGLE ; GAO YING ; WU HONGRUI ; CHEN JIXIANG</creator><creatorcontrib>MIAO HONGLE ; GAO YING ; WU HONGRUI ; CHEN JIXIANG</creatorcontrib><description>The invention discloses a model security detection method based on a generative adversarial network. The method comprises the following specific steps of: 1, setting a behavior similarity security threshold value delta; 2, constructing an initialization generator G and a substitution model D; 3, carrying out iteration round by round, wherein the behavior similarity mu between the substitution model D and the tested model T is calculated, and the next step is carried out after the behavior similarity mu reaches a set value; and 4, evaluating the safety of the tested model T. The model stealing method based on the GAN is suitable for a black box attack scene without training data, and the behavior similarity of the substitution model and the tested model is rapidly improved by generating artificial data with relatively balanced categories. According to experimental results, the method has the characteristics of high adaptability, high efficiency and the like, and the risk of model stealing attack after the acce</description><language>chi ; eng</language><subject>CALCULATING ; COMPUTING ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; PHYSICS</subject><creationdate>2021</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=20210618&amp;DB=EPODOC&amp;CC=CN&amp;NR=112989361A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25564,76547</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20210618&amp;DB=EPODOC&amp;CC=CN&amp;NR=112989361A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>MIAO HONGLE</creatorcontrib><creatorcontrib>GAO YING</creatorcontrib><creatorcontrib>WU HONGRUI</creatorcontrib><creatorcontrib>CHEN JIXIANG</creatorcontrib><title>Model security detection method based on generative adversarial network</title><description>The invention discloses a model security detection method based on a generative adversarial network. The method comprises the following specific steps of: 1, setting a behavior similarity security threshold value delta; 2, constructing an initialization generator G and a substitution model D; 3, carrying out iteration round by round, wherein the behavior similarity mu between the substitution model D and the tested model T is calculated, and the next step is carried out after the behavior similarity mu reaches a set value; and 4, evaluating the safety of the tested model T. The model stealing method based on the GAN is suitable for a black box attack scene without training data, and the behavior similarity of the substitution model and the tested model is rapidly improved by generating artificial data with relatively balanced categories. According to experimental results, the method has the characteristics of high adaptability, high efficiency and the like, and the risk of model stealing attack after the acce</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>2021</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNrjZHD3zU9JzVEoTk0uLcosqVRISS1JTS7JzM9TyE0tychPUUhKLE5NUQDy01PzUosSSzLLUhUSU8pSi4oTizITcxTyUkvK84uyeRhY0xJzilN5oTQ3g6Kba4izh25qQX58anFBYjJQe0m8s5-hoZGlhaWxmaGjMTFqAKRoNIU</recordid><startdate>20210618</startdate><enddate>20210618</enddate><creator>MIAO HONGLE</creator><creator>GAO YING</creator><creator>WU HONGRUI</creator><creator>CHEN JIXIANG</creator><scope>EVB</scope></search><sort><creationdate>20210618</creationdate><title>Model security detection method based on generative adversarial network</title><author>MIAO HONGLE ; GAO YING ; WU HONGRUI ; CHEN JIXIANG</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_CN112989361A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>chi ; eng</language><creationdate>2021</creationdate><topic>CALCULATING</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>MIAO HONGLE</creatorcontrib><creatorcontrib>GAO YING</creatorcontrib><creatorcontrib>WU HONGRUI</creatorcontrib><creatorcontrib>CHEN JIXIANG</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>MIAO HONGLE</au><au>GAO YING</au><au>WU HONGRUI</au><au>CHEN JIXIANG</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Model security detection method based on generative adversarial network</title><date>2021-06-18</date><risdate>2021</risdate><abstract>The invention discloses a model security detection method based on a generative adversarial network. The method comprises the following specific steps of: 1, setting a behavior similarity security threshold value delta; 2, constructing an initialization generator G and a substitution model D; 3, carrying out iteration round by round, wherein the behavior similarity mu between the substitution model D and the tested model T is calculated, and the next step is carried out after the behavior similarity mu reaches a set value; and 4, evaluating the safety of the tested model T. The model stealing method based on the GAN is suitable for a black box attack scene without training data, and the behavior similarity of the substitution model and the tested model is rapidly improved by generating artificial data with relatively balanced categories. According to experimental results, the method has the characteristics of high adaptability, high efficiency and the like, and the risk of model stealing attack after the acce</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language chi ; eng
recordid cdi_epo_espacenet_CN112989361A
source esp@cenet
subjects CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
title Model security detection method based on generative adversarial network
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T06%3A26%3A21IST&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=MIAO%20HONGLE&rft.date=2021-06-18&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3ECN112989361A%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