A Causal Framework for Decomposing Spurious Variations

One of the fundamental challenges found throughout the data sciences is to explain why things happen in specific ways, or through which mechanisms a certain variable $X$ exerts influences over another variable $Y$. In statistics and machine learning, significant efforts have been put into developing...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Plecko, Drago, Bareinboim, Elias
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 Plecko, Drago
Bareinboim, Elias
description One of the fundamental challenges found throughout the data sciences is to explain why things happen in specific ways, or through which mechanisms a certain variable $X$ exerts influences over another variable $Y$. In statistics and machine learning, significant efforts have been put into developing machinery to estimate correlations across variables efficiently. In causal inference, a large body of literature is concerned with the decomposition of causal effects under the rubric of mediation analysis. However, many variations are spurious in nature, including different phenomena throughout the applied sciences. Despite the statistical power to estimate correlations and the identification power to decompose causal effects, there is still little understanding of the properties of spurious associations and how they can be decomposed in terms of the underlying causal mechanisms. In this manuscript, we develop formal tools for decomposing spurious variations in both Markovian and Semi-Markovian models. We prove the first results that allow a non-parametric decomposition of spurious effects and provide sufficient conditions for the identification of such decompositions. The described approach has several applications, ranging from explainable and fair AI to questions in epidemiology and medicine, and we empirically demonstrate its use on a real-world dataset.
doi_str_mv 10.48550/arxiv.2306.05071
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2306_05071</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2306_05071</sourcerecordid><originalsourceid>FETCH-LOGICAL-a671-6d79bc0b6195e77a017e372780f4a54bfa8d86c38d83d1b666c8f1bf85d7c08f3</originalsourceid><addsrcrecordid>eNotj7FOwzAURb0woMIHMOEfSHiuYz93rEILSJU6tOoaPTs2smjqyCYF_h4oLOduR_cwdiegboxS8ED5M57ruQRdgwIU10wveUtToSNfZxr8R8pvPKTMH71Lw5hKPL3y3TjlmKbCD5Qjvcd0KjfsKtCx-Nv_nbH9erVvn6vN9umlXW4q0igq3ePCOrBaLJRHJBDoJc7RQGhINTaQ6Y128oeyF1Zr7UwQNhjVowMT5Izd_2kvx7sxx4HyV_cb0F0C5DfeqEAb</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>A Causal Framework for Decomposing Spurious Variations</title><source>arXiv.org</source><creator>Plecko, Drago ; Bareinboim, Elias</creator><creatorcontrib>Plecko, Drago ; Bareinboim, Elias</creatorcontrib><description>One of the fundamental challenges found throughout the data sciences is to explain why things happen in specific ways, or through which mechanisms a certain variable $X$ exerts influences over another variable $Y$. In statistics and machine learning, significant efforts have been put into developing machinery to estimate correlations across variables efficiently. In causal inference, a large body of literature is concerned with the decomposition of causal effects under the rubric of mediation analysis. However, many variations are spurious in nature, including different phenomena throughout the applied sciences. Despite the statistical power to estimate correlations and the identification power to decompose causal effects, there is still little understanding of the properties of spurious associations and how they can be decomposed in terms of the underlying causal mechanisms. In this manuscript, we develop formal tools for decomposing spurious variations in both Markovian and Semi-Markovian models. We prove the first results that allow a non-parametric decomposition of spurious effects and provide sufficient conditions for the identification of such decompositions. The described approach has several applications, ranging from explainable and fair AI to questions in epidemiology and medicine, and we empirically demonstrate its use on a real-world dataset.</description><identifier>DOI: 10.48550/arxiv.2306.05071</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Learning ; Statistics - Machine Learning ; Statistics - Methodology</subject><creationdate>2023-06</creationdate><rights>http://creativecommons.org/licenses/by/4.0</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/2306.05071$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2306.05071$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Plecko, Drago</creatorcontrib><creatorcontrib>Bareinboim, Elias</creatorcontrib><title>A Causal Framework for Decomposing Spurious Variations</title><description>One of the fundamental challenges found throughout the data sciences is to explain why things happen in specific ways, or through which mechanisms a certain variable $X$ exerts influences over another variable $Y$. In statistics and machine learning, significant efforts have been put into developing machinery to estimate correlations across variables efficiently. In causal inference, a large body of literature is concerned with the decomposition of causal effects under the rubric of mediation analysis. However, many variations are spurious in nature, including different phenomena throughout the applied sciences. Despite the statistical power to estimate correlations and the identification power to decompose causal effects, there is still little understanding of the properties of spurious associations and how they can be decomposed in terms of the underlying causal mechanisms. In this manuscript, we develop formal tools for decomposing spurious variations in both Markovian and Semi-Markovian models. We prove the first results that allow a non-parametric decomposition of spurious effects and provide sufficient conditions for the identification of such decompositions. The described approach has several applications, ranging from explainable and fair AI to questions in epidemiology and medicine, and we empirically demonstrate its use on a real-world dataset.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Learning</subject><subject>Statistics - Machine Learning</subject><subject>Statistics - Methodology</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj7FOwzAURb0woMIHMOEfSHiuYz93rEILSJU6tOoaPTs2smjqyCYF_h4oLOduR_cwdiegboxS8ED5M57ruQRdgwIU10wveUtToSNfZxr8R8pvPKTMH71Lw5hKPL3y3TjlmKbCD5Qjvcd0KjfsKtCx-Nv_nbH9erVvn6vN9umlXW4q0igq3ePCOrBaLJRHJBDoJc7RQGhINTaQ6Y128oeyF1Zr7UwQNhjVowMT5Izd_2kvx7sxx4HyV_cb0F0C5DfeqEAb</recordid><startdate>20230608</startdate><enddate>20230608</enddate><creator>Plecko, Drago</creator><creator>Bareinboim, Elias</creator><scope>AKY</scope><scope>EPD</scope><scope>GOX</scope></search><sort><creationdate>20230608</creationdate><title>A Causal Framework for Decomposing Spurious Variations</title><author>Plecko, Drago ; Bareinboim, Elias</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a671-6d79bc0b6195e77a017e372780f4a54bfa8d86c38d83d1b666c8f1bf85d7c08f3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Learning</topic><topic>Statistics - Machine Learning</topic><topic>Statistics - Methodology</topic><toplevel>online_resources</toplevel><creatorcontrib>Plecko, Drago</creatorcontrib><creatorcontrib>Bareinboim, Elias</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv Statistics</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Plecko, Drago</au><au>Bareinboim, Elias</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>A Causal Framework for Decomposing Spurious Variations</atitle><date>2023-06-08</date><risdate>2023</risdate><abstract>One of the fundamental challenges found throughout the data sciences is to explain why things happen in specific ways, or through which mechanisms a certain variable $X$ exerts influences over another variable $Y$. In statistics and machine learning, significant efforts have been put into developing machinery to estimate correlations across variables efficiently. In causal inference, a large body of literature is concerned with the decomposition of causal effects under the rubric of mediation analysis. However, many variations are spurious in nature, including different phenomena throughout the applied sciences. Despite the statistical power to estimate correlations and the identification power to decompose causal effects, there is still little understanding of the properties of spurious associations and how they can be decomposed in terms of the underlying causal mechanisms. In this manuscript, we develop formal tools for decomposing spurious variations in both Markovian and Semi-Markovian models. We prove the first results that allow a non-parametric decomposition of spurious effects and provide sufficient conditions for the identification of such decompositions. The described approach has several applications, ranging from explainable and fair AI to questions in epidemiology and medicine, and we empirically demonstrate its use on a real-world dataset.</abstract><doi>10.48550/arxiv.2306.05071</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2306.05071
ispartof
issn
language eng
recordid cdi_arxiv_primary_2306_05071
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Learning
Statistics - Machine Learning
Statistics - Methodology
title A Causal Framework for Decomposing Spurious Variations
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-18T05%3A57%3A21IST&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=A%20Causal%20Framework%20for%20Decomposing%20Spurious%20Variations&rft.au=Plecko,%20Drago&rft.date=2023-06-08&rft_id=info:doi/10.48550/arxiv.2306.05071&rft_dat=%3Carxiv_GOX%3E2306_05071%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