Risk mitigation in algorithmic accountability: The role of machine learning copies

Machine learning plays an increasingly important role in our society and economy and is already having an impact on our daily life in many different ways. From several perspectives, machine learning is seen as the new engine of productivity and economic growth. It can increase the business efficienc...

Descripción completa

Detalles Bibliográficos
Autores: Unceta, Irene, Pujol, Oriol, Nin, Jordi
Tipo de recurso: artículo
Fecha de publicación:2020
País:España
Institución:Universitat Ramon Llull (URL)
Repositorio:DAU Arxiu Digital de la Universitat Ramon Llull
OAI Identifier:oai:dau.url.edu:20.500.14342/5086
Acceso en línea:http://hdl.handle.net/20.500.14342/5086
http://doi.org/10.1371/journal.pone.0241286
Access Level:acceso abierto
Palabra clave:Machine learning systems
id ES_bd62a54b2fca2e24deb0d35a35c71d82
oai_identifier_str oai:dau.url.edu:20.500.14342/5086
network_acronym_str ES
network_name_str España
repository_id_str
spelling Risk mitigation in algorithmic accountability: The role of machine learning copiesUnceta, IrenePujol, OriolNin, JordiMachine learning systemsMachine learning plays an increasingly important role in our society and economy and is already having an impact on our daily life in many different ways. From several perspectives, machine learning is seen as the new engine of productivity and economic growth. It can increase the business efficiency and improve any decision-making process, and of course, spawn the creation of new products and services by using complex machine learning algorithms. In this scenario, the lack of actionable accountability-related guidance is potentially the single most important challenge facing the machine learning community. Machine learning systems are often composed of many parts and ingredients, mixing third party components or software-as-a-service APIs, among others. In this paper we study the role of copies for risk mitigation in such machine learning systems. Formally, a copy can be regarded as an approximated projection operator of a model into a target model hypothesis set. Under the conceptual framework of actionable accountability, we explore the use of copies as a viable alternative in circumstances where models cannot be re-trained, nor enhanced by means of a wrapper. We use a real residential mortgage default dataset as a use case to illustrate the feasibility of this approach.info:eu-repo/semantics/publishedVersionPublic Library of ScienceUniversitat Ramon Llull. Esade202520252020info:eu-repo/semantics/article26 p.application/pdfhttp://hdl.handle.net/20.500.14342/5086http://doi.org/10.1371/journal.pone.0241286reponame:DAU Arxiu Digital de la Universitat Ramon Llullinstname:Universitat Ramon Llull (URL)InglésPLOS One© L'autor/aAttribution 4.0 Internationalhttp://creativecommons.org/licenses/by/4.0/info:eu-repo/semantics/openAccessoai:dau.url.edu:20.500.14342/50862026-06-21T06:40:37Z
dc.title.none.fl_str_mv Risk mitigation in algorithmic accountability: The role of machine learning copies
title Risk mitigation in algorithmic accountability: The role of machine learning copies
spellingShingle Risk mitigation in algorithmic accountability: The role of machine learning copies
Unceta, Irene
Machine learning systems
title_short Risk mitigation in algorithmic accountability: The role of machine learning copies
title_full Risk mitigation in algorithmic accountability: The role of machine learning copies
title_fullStr Risk mitigation in algorithmic accountability: The role of machine learning copies
title_full_unstemmed Risk mitigation in algorithmic accountability: The role of machine learning copies
title_sort Risk mitigation in algorithmic accountability: The role of machine learning copies
dc.creator.none.fl_str_mv Unceta, Irene
Pujol, Oriol
Nin, Jordi
author Unceta, Irene
author_facet Unceta, Irene
Pujol, Oriol
Nin, Jordi
author_role author
author2 Pujol, Oriol
Nin, Jordi
author2_role author
author
dc.contributor.none.fl_str_mv Universitat Ramon Llull. Esade
dc.subject.none.fl_str_mv Machine learning systems
topic Machine learning systems
description Machine learning plays an increasingly important role in our society and economy and is already having an impact on our daily life in many different ways. From several perspectives, machine learning is seen as the new engine of productivity and economic growth. It can increase the business efficiency and improve any decision-making process, and of course, spawn the creation of new products and services by using complex machine learning algorithms. In this scenario, the lack of actionable accountability-related guidance is potentially the single most important challenge facing the machine learning community. Machine learning systems are often composed of many parts and ingredients, mixing third party components or software-as-a-service APIs, among others. In this paper we study the role of copies for risk mitigation in such machine learning systems. Formally, a copy can be regarded as an approximated projection operator of a model into a target model hypothesis set. Under the conceptual framework of actionable accountability, we explore the use of copies as a viable alternative in circumstances where models cannot be re-trained, nor enhanced by means of a wrapper. We use a real residential mortgage default dataset as a use case to illustrate the feasibility of this approach.
publishDate 2020
dc.date.none.fl_str_mv 2020
2025
2025
dc.type.none.fl_str_mv info:eu-repo/semantics/article
format article
dc.identifier.none.fl_str_mv http://hdl.handle.net/20.500.14342/5086
http://doi.org/10.1371/journal.pone.0241286
url http://hdl.handle.net/20.500.14342/5086
http://doi.org/10.1371/journal.pone.0241286
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.relation.none.fl_str_mv PLOS One
dc.rights.none.fl_str_mv © L'autor/a
Attribution 4.0 International
http://creativecommons.org/licenses/by/4.0/
info:eu-repo/semantics/openAccess
rights_invalid_str_mv © L'autor/a
Attribution 4.0 International
http://creativecommons.org/licenses/by/4.0/
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv 26 p.
application/pdf
dc.publisher.none.fl_str_mv Public Library of Science
publisher.none.fl_str_mv Public Library of Science
dc.source.none.fl_str_mv reponame:DAU Arxiu Digital de la Universitat Ramon Llull
instname:Universitat Ramon Llull (URL)
instname_str Universitat Ramon Llull (URL)
reponame_str DAU Arxiu Digital de la Universitat Ramon Llull
collection DAU Arxiu Digital de la Universitat Ramon Llull
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869418196006600704
score 15.811543

Similares en LA Referencia