Understanding surrogate explanations: the interplay between complexity, fidelity and coverage

This paper analyses the fundamental ingredients behind surrogate explanations to provide a better understanding of their inner workings. We start our exposition by considering global surrogates, describing the trade-off between complexity of the surrogate and fidelity to the black-box being modelled...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Poyiadzi, Rafael, Renard, Xavier, Laugel, Thibault, Santos-Rodriguez, Raul, Detyniecki, Marcin
Format: Artikel
Sprache: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 Poyiadzi, Rafael
Renard, Xavier
Laugel, Thibault
Santos-Rodriguez, Raul
Detyniecki, Marcin
description This paper analyses the fundamental ingredients behind surrogate explanations to provide a better understanding of their inner workings. We start our exposition by considering global surrogates, describing the trade-off between complexity of the surrogate and fidelity to the black-box being modelled. We show that transitioning from global to local - reducing coverage - allows for more favourable conditions on the Pareto frontier of fidelity-complexity of a surrogate. We discuss the interplay between complexity, fidelity and coverage, and consider how different user needs can lead to problem formulations where these are either constraints or penalties. We also present experiments that demonstrate how the local surrogate interpretability procedure can be made interactive and lead to better explanations.
doi_str_mv 10.48550/arxiv.2107.04309
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2107_04309</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2107_04309</sourcerecordid><originalsourceid>FETCH-LOGICAL-a679-c71688163dd2e8e3e369a085b8057d011e7b5bc779ffbccad9e6a8558d35bf0c3</originalsourceid><addsrcrecordid>eNotj8tOwzAURL1hgQofwAp_AAl2XccOO1TxkiqxKUsUXds3wVLqRI4pyd9jSlczmpGOdAi54azcaCnZPcTZH8s1Z6pkG8HqS_L5ERzGKUFwPnR0-o5x6CAhxXnsIUDyQ5geaPpC6kPCmMeFGkw_iIHa4TD2OPu03NHWO-xzo5mUjyNG6PCKXLTQT3h9zhXZPz_tt6_F7v3lbfu4K6BSdWEVr7TmlXBujRoFiqoGpqXRTCrHOEdlpLFK1W1rrAVXYwXZRzshTcusWJHbf-zJrxmjP0Bcmj_P5uQpfgEBP1BM</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Understanding surrogate explanations: the interplay between complexity, fidelity and coverage</title><source>arXiv.org</source><creator>Poyiadzi, Rafael ; Renard, Xavier ; Laugel, Thibault ; Santos-Rodriguez, Raul ; Detyniecki, Marcin</creator><creatorcontrib>Poyiadzi, Rafael ; Renard, Xavier ; Laugel, Thibault ; Santos-Rodriguez, Raul ; Detyniecki, Marcin</creatorcontrib><description>This paper analyses the fundamental ingredients behind surrogate explanations to provide a better understanding of their inner workings. We start our exposition by considering global surrogates, describing the trade-off between complexity of the surrogate and fidelity to the black-box being modelled. We show that transitioning from global to local - reducing coverage - allows for more favourable conditions on the Pareto frontier of fidelity-complexity of a surrogate. We discuss the interplay between complexity, fidelity and coverage, and consider how different user needs can lead to problem formulations where these are either constraints or penalties. We also present experiments that demonstrate how the local surrogate interpretability procedure can be made interactive and lead to better explanations.</description><identifier>DOI: 10.48550/arxiv.2107.04309</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Learning</subject><creationdate>2021-07</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,777,882</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2107.04309$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2107.04309$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Poyiadzi, Rafael</creatorcontrib><creatorcontrib>Renard, Xavier</creatorcontrib><creatorcontrib>Laugel, Thibault</creatorcontrib><creatorcontrib>Santos-Rodriguez, Raul</creatorcontrib><creatorcontrib>Detyniecki, Marcin</creatorcontrib><title>Understanding surrogate explanations: the interplay between complexity, fidelity and coverage</title><description>This paper analyses the fundamental ingredients behind surrogate explanations to provide a better understanding of their inner workings. We start our exposition by considering global surrogates, describing the trade-off between complexity of the surrogate and fidelity to the black-box being modelled. We show that transitioning from global to local - reducing coverage - allows for more favourable conditions on the Pareto frontier of fidelity-complexity of a surrogate. We discuss the interplay between complexity, fidelity and coverage, and consider how different user needs can lead to problem formulations where these are either constraints or penalties. We also present experiments that demonstrate how the local surrogate interpretability procedure can be made interactive and lead to better explanations.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj8tOwzAURL1hgQofwAp_AAl2XccOO1TxkiqxKUsUXds3wVLqRI4pyd9jSlczmpGOdAi54azcaCnZPcTZH8s1Z6pkG8HqS_L5ERzGKUFwPnR0-o5x6CAhxXnsIUDyQ5geaPpC6kPCmMeFGkw_iIHa4TD2OPu03NHWO-xzo5mUjyNG6PCKXLTQT3h9zhXZPz_tt6_F7v3lbfu4K6BSdWEVr7TmlXBujRoFiqoGpqXRTCrHOEdlpLFK1W1rrAVXYwXZRzshTcusWJHbf-zJrxmjP0Bcmj_P5uQpfgEBP1BM</recordid><startdate>20210709</startdate><enddate>20210709</enddate><creator>Poyiadzi, Rafael</creator><creator>Renard, Xavier</creator><creator>Laugel, Thibault</creator><creator>Santos-Rodriguez, Raul</creator><creator>Detyniecki, Marcin</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20210709</creationdate><title>Understanding surrogate explanations: the interplay between complexity, fidelity and coverage</title><author>Poyiadzi, Rafael ; Renard, Xavier ; Laugel, Thibault ; Santos-Rodriguez, Raul ; Detyniecki, Marcin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a679-c71688163dd2e8e3e369a085b8057d011e7b5bc779ffbccad9e6a8558d35bf0c3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Poyiadzi, Rafael</creatorcontrib><creatorcontrib>Renard, Xavier</creatorcontrib><creatorcontrib>Laugel, Thibault</creatorcontrib><creatorcontrib>Santos-Rodriguez, Raul</creatorcontrib><creatorcontrib>Detyniecki, Marcin</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Poyiadzi, Rafael</au><au>Renard, Xavier</au><au>Laugel, Thibault</au><au>Santos-Rodriguez, Raul</au><au>Detyniecki, Marcin</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Understanding surrogate explanations: the interplay between complexity, fidelity and coverage</atitle><date>2021-07-09</date><risdate>2021</risdate><abstract>This paper analyses the fundamental ingredients behind surrogate explanations to provide a better understanding of their inner workings. We start our exposition by considering global surrogates, describing the trade-off between complexity of the surrogate and fidelity to the black-box being modelled. We show that transitioning from global to local - reducing coverage - allows for more favourable conditions on the Pareto frontier of fidelity-complexity of a surrogate. We discuss the interplay between complexity, fidelity and coverage, and consider how different user needs can lead to problem formulations where these are either constraints or penalties. We also present experiments that demonstrate how the local surrogate interpretability procedure can be made interactive and lead to better explanations.</abstract><doi>10.48550/arxiv.2107.04309</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2107.04309
ispartof
issn
language eng
recordid cdi_arxiv_primary_2107_04309
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Learning
title Understanding surrogate explanations: the interplay between complexity, fidelity and coverage
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-20T10%3A39%3A36IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Understanding%20surrogate%20explanations:%20the%20interplay%20between%20complexity,%20fidelity%20and%20coverage&rft.au=Poyiadzi,%20Rafael&rft.date=2021-07-09&rft_id=info:doi/10.48550/arxiv.2107.04309&rft_dat=%3Carxiv_GOX%3E2107_04309%3C/arxiv_GOX%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