国际清算银行-管理解释：监管机构如何解决人工智能的可解释性（英）-2025.9-23页.pdf

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。

OccasionalPaper

No24

Managingexplanations:

howregulatorscan

addressAIexplainability

byFernandoPérez-Cruz,JermyPrenio,Fernando

Restoy,JefferyYong

September2025

JELclassification:C60,G29,G38,O30

Keywords:artificialintelligence,machinelearning,

modelriskmanagement,riskgovernance

FSIOccasionalPapersaimtocontributetointernationaldiscussionsonawiderangeoftopicsofrelevance

tothefinancialindustryanditsregulationandsupervision.Theviewsexpressedinthemaresolelythose

oftheauthorsanddonotnecessarilyreflectthoseoftheBISortheBasel-basedstandard-settingbodies.

ThispublicationisavailableontheBISwebsite().TocontacttheBISGlobalMediaandPublic

Relationsteam,pleaseemailmedia@.Youcansignupforemailalertsat

/emailalerts.htm.

©BankforInternationalSettlements2025.Allrightsreserved.Briefexcerptsmaybereproducedor

translatedprovidedthesourceisstated.

ISSN1020-9999(online)

Abstract

Theincreasingadoptionofartificialintelligence(AI)byfinancialinstitutionsistransformingtheir

operations,riskmanagementandcustomerinteractions.Nevertheless,thelimitedexplainabilityof

complexAImodels,particularlywhenusedincriticalbusinessapplications,posessignificantchallenges

andissuesforfinancialinstitutionsandregulators.Explainability,ortheextenttowhichamodel’soutput

canbeexplainedtoahuman,isessentialfortransparency,accountability,regulatorycomplianceand

consumertrust.Yet,complexAImodels,suchasdeeplearningandlargelanguagemodels(LLMs),are

oftendifficulttoexplain.Whilethereareexistingexplainabilitytechniquesthatcanhelpshedlighton

complexAImodels’behaviour,thes

该用户很懒，什么也没介绍

咨询Ta 进入空间

更多 >