Scalable Out-of-distribution Robustness in the Presence of Unobserved Confounders Article Swipe

PDF

Parjanya Prashant , Seyedeh Baharan Khatami , Bruno Ribeiro , Babak Salimi ·

YOU? · · 2024 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2411.19923

We consider the task of out-of-distribution (OOD) generalization, where the distribution shift is due to an unobserved confounder ($Z$) affecting both the covariates ($X$) and the labels ($Y$). This confounding introduces heterogeneity in the predictor, i.e., $P(Y | X) = E_{P(Z | X)}[P(Y | X,Z)]$, making traditional covariate and label shift assumptions unsuitable. OOD generalization differs from traditional domain adaptation in that it does not assume access to the covariate distribution ($X^\text{te}$) of the test samples during training. These conditions create a challenging scenario for OOD robustness: (a) $Z^\text{tr}$ is an unobserved confounder during training, (b) $P^\text{te}(Z) \neq P^\text{tr}(Z)$, (c) $X^\text{te}$ is unavailable during training, and (d) the predictive distribution depends on $P^\text{te}(Z)$. While prior work has developed complex predictors requiring multiple additional variables for identifiability of the latent distribution, we explore a set of identifiability assumptions that yield a surprisingly simple predictor using only a single additional variable. Our approach demonstrates superior empirical performance on several benchmark tasks.

Related Topics

Concepts

Robustness (evolution) Confounding Scalability Econometrics Computer science Statistics Mathematics Chemistry Database Gene Biochemistry

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2411.19923
PDF: https://arxiv.org/pdf/2411.19923
OA Status: green
Related Works: 10
OpenAlex ID: https://openalex.org/W4405031615

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4405031615

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2411.19923

Digital Object Identifier
Title: Scalable Out-of-distribution Robustness in the Presence of Unobserved Confounders

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2024

Year of publication
Publication date: 2024-11-29

Full publication date if available
Authors: Parjanya Prashant, Seyedeh Baharan Khatami, Bruno Ribeiro, Babak Salimi

List of authors in order
Landing page: https://arxiv.org/abs/2411.19923

Publisher landing page
PDF URL: https://arxiv.org/pdf/2411.19923

Direct link to full text PDF
Open access: Yes

Whether a free full text is available
OA status: green

Open access status per OpenAlex
OA URL: https://arxiv.org/pdf/2411.19923

Direct OA link when available
Concepts: Robustness (evolution), Confounding, Scalability, Econometrics, Computer science, Statistics, Mathematics, Chemistry, Database, Gene, Biochemistry

Top concepts (fields/topics) attached by OpenAlex
Cited by: 0

Total citation count in OpenAlex
Related works (count): 10

Other works algorithmically related by OpenAlex

Full payload

id	https://openalex.org/W4405031615
doi	https://doi.org/10.48550/arxiv.2411.19923
ids.doi	https://doi.org/10.48550/arxiv.2411.19923
ids.openalex	https://openalex.org/W4405031615
fwci
type	preprint
title	Scalable Out-of-distribution Robustness in the Presence of Unobserved Confounders
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
topics[0].id	https://openalex.org/T11413
topics[0].field.id	https://openalex.org/fields/18
topics[0].field.display_name	Decision Sciences
topics[0].score	0.7652999758720398
topics[0].domain.id	https://openalex.org/domains/2
topics[0].domain.display_name	Social Sciences
topics[0].subfield.id	https://openalex.org/subfields/1803
topics[0].subfield.display_name	Management Science and Operations Research
topics[0].display_name	Risk and Portfolio Optimization
is_xpac	False
apc_list
apc_paid
concepts[0].id	https://openalex.org/C63479239
concepts[0].level	3
concepts[0].score	0.8038744926452637
concepts[0].wikidata	https://www.wikidata.org/wiki/Q7353546
concepts[0].display_name	Robustness (evolution)
concepts[1].id	https://openalex.org/C77350462
concepts[1].level	2
concepts[1].score	0.7467128038406372
concepts[1].wikidata	https://www.wikidata.org/wiki/Q1125472
concepts[1].display_name	Confounding
concepts[2].id	https://openalex.org/C48044578
concepts[2].level	2
concepts[2].score	0.5718826651573181
concepts[2].wikidata	https://www.wikidata.org/wiki/Q727490
concepts[2].display_name	Scalability
concepts[3].id	https://openalex.org/C149782125
concepts[3].level	1
concepts[3].score	0.563908040523529
concepts[3].wikidata	https://www.wikidata.org/wiki/Q160039
concepts[3].display_name	Econometrics
concepts[4].id	https://openalex.org/C41008148
concepts[4].level	0
concepts[4].score	0.49621516466140747
concepts[4].wikidata	https://www.wikidata.org/wiki/Q21198
concepts[4].display_name	Computer science
concepts[5].id	https://openalex.org/C105795698
concepts[5].level	1
concepts[5].score	0.42319193482398987
concepts[5].wikidata	https://www.wikidata.org/wiki/Q12483
concepts[5].display_name	Statistics
concepts[6].id	https://openalex.org/C33923547
concepts[6].level	0
concepts[6].score	0.2959364652633667
concepts[6].wikidata	https://www.wikidata.org/wiki/Q395
concepts[6].display_name	Mathematics
concepts[7].id	https://openalex.org/C185592680
concepts[7].level	0
concepts[7].score	0.11424347758293152
concepts[7].wikidata	https://www.wikidata.org/wiki/Q2329
concepts[7].display_name	Chemistry
concepts[8].id	https://openalex.org/C77088390
concepts[8].level	1
concepts[8].score	0.09276077151298523
concepts[8].wikidata	https://www.wikidata.org/wiki/Q8513
concepts[8].display_name	Database
concepts[9].id	https://openalex.org/C104317684
concepts[9].level	2
concepts[9].score	0.0
concepts[9].wikidata	https://www.wikidata.org/wiki/Q7187
concepts[9].display_name	Gene
concepts[10].id	https://openalex.org/C55493867
concepts[10].level	1
concepts[10].score	0.0
concepts[10].wikidata	https://www.wikidata.org/wiki/Q7094
concepts[10].display_name	Biochemistry
keywords[0].id	https://openalex.org/keywords/robustness
keywords[0].score	0.8038744926452637
keywords[0].display_name	Robustness (evolution)
keywords[1].id	https://openalex.org/keywords/confounding
keywords[1].score	0.7467128038406372
keywords[1].display_name	Confounding
keywords[2].id	https://openalex.org/keywords/scalability
keywords[2].score	0.5718826651573181
keywords[2].display_name	Scalability
keywords[3].id	https://openalex.org/keywords/econometrics
keywords[3].score	0.563908040523529
keywords[3].display_name	Econometrics
keywords[4].id	https://openalex.org/keywords/computer-science
keywords[4].score	0.49621516466140747
keywords[4].display_name	Computer science
keywords[5].id	https://openalex.org/keywords/statistics
keywords[5].score	0.42319193482398987
keywords[5].display_name	Statistics
keywords[6].id	https://openalex.org/keywords/mathematics
keywords[6].score	0.2959364652633667
keywords[6].display_name	Mathematics
keywords[7].id	https://openalex.org/keywords/chemistry
keywords[7].score	0.11424347758293152
keywords[7].display_name	Chemistry
keywords[8].id	https://openalex.org/keywords/database
keywords[8].score	0.09276077151298523
keywords[8].display_name	Database
language	en
locations[0].id	pmh:oai:arXiv.org:2411.19923
locations[0].is_oa	True
locations[0].source.id	https://openalex.org/S4306400194
locations[0].source.issn
locations[0].source.type	repository
locations[0].source.is_oa	True
locations[0].source.issn_l
locations[0].source.is_core	False
locations[0].source.is_in_doaj	False
locations[0].source.display_name	arXiv (Cornell University)
locations[0].source.host_organization	https://openalex.org/I205783295
locations[0].source.host_organization_name	Cornell University
locations[0].source.host_organization_lineage	https://openalex.org/I205783295
locations[0].license
locations[0].pdf_url	https://arxiv.org/pdf/2411.19923
locations[0].version	submittedVersion
locations[0].raw_type	text
locations[0].license_id
locations[0].is_accepted	False
locations[0].is_published	False
locations[0].raw_source_name
locations[0].landing_page_url	http://arxiv.org/abs/2411.19923
locations[1].id	doi:10.48550/arxiv.2411.19923
locations[1].is_oa	True
locations[1].source.id	https://openalex.org/S4306400194
locations[1].source.issn
locations[1].source.type	repository
locations[1].source.is_oa	True
locations[1].source.issn_l
locations[1].source.is_core	False
locations[1].source.is_in_doaj	False
locations[1].source.display_name	arXiv (Cornell University)
locations[1].source.host_organization	https://openalex.org/I205783295
locations[1].source.host_organization_name	Cornell University
locations[1].source.host_organization_lineage	https://openalex.org/I205783295
locations[1].license
locations[1].pdf_url
locations[1].version
locations[1].raw_type	article
locations[1].license_id
locations[1].is_accepted	False
locations[1].is_published
locations[1].raw_source_name
locations[1].landing_page_url	https://doi.org/10.48550/arxiv.2411.19923
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5082514869
authorships[0].author.orcid	https://orcid.org/0000-0003-1718-3037
authorships[0].author.display_name	Parjanya Prashant
authorships[0].author_position	first
authorships[0].raw_author_name	Prashant, Parjanya
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5094196472
authorships[1].author.orcid
authorships[1].author.display_name	Seyedeh Baharan Khatami
authorships[1].author_position	middle
authorships[1].raw_author_name	Khatami, Seyedeh Baharan
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5076743302
authorships[2].author.orcid	https://orcid.org/0000-0001-5105-8905
authorships[2].author.display_name	Bruno Ribeiro
authorships[2].author_position	middle
authorships[2].raw_author_name	Ribeiro, Bruno
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5103209063
authorships[3].author.orcid	https://orcid.org/0000-0003-2485-9533
authorships[3].author.display_name	Babak Salimi
authorships[3].author_position	last
authorships[3].raw_author_name	Salimi, Babak
authorships[3].is_corresponding	False
has_content.pdf	False
has_content.grobid_xml	False
is_paratext	False
open_access.is_oa	True
open_access.oa_url	https://arxiv.org/pdf/2411.19923
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-10-10T00:00:00
display_name	Scalable Out-of-distribution Robustness in the Presence of Unobserved Confounders
has_fulltext	False
is_retracted	False
updated_date	2025-11-06T06:51:31.235846
primary_topic.id	https://openalex.org/T11413
primary_topic.field.id	https://openalex.org/fields/18
primary_topic.field.display_name	Decision Sciences
primary_topic.score	0.7652999758720398
primary_topic.domain.id	https://openalex.org/domains/2
primary_topic.domain.display_name	Social Sciences
primary_topic.subfield.id	https://openalex.org/subfields/1803
primary_topic.subfield.display_name	Management Science and Operations Research
primary_topic.display_name	Risk and Portfolio Optimization
related_works	https://openalex.org/W2994176440, https://openalex.org/W2510575233, https://openalex.org/W2481749367, https://openalex.org/W830718730, https://openalex.org/W2495367848, https://openalex.org/W2793477322, https://openalex.org/W4236720793, https://openalex.org/W1992228662, https://openalex.org/W4246786946, https://openalex.org/W2503931704
cited_by_count	0
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2411.19923
best_oa_location.is_oa	True
best_oa_location.source.id	https://openalex.org/S4306400194
best_oa_location.source.issn
best_oa_location.source.type	repository
best_oa_location.source.is_oa	True
best_oa_location.source.issn_l
best_oa_location.source.is_core	False
best_oa_location.source.is_in_doaj	False
best_oa_location.source.display_name	arXiv (Cornell University)
best_oa_location.source.host_organization	https://openalex.org/I205783295
best_oa_location.source.host_organization_name	Cornell University
best_oa_location.source.host_organization_lineage	https://openalex.org/I205783295
best_oa_location.license
best_oa_location.pdf_url	https://arxiv.org/pdf/2411.19923
best_oa_location.version	submittedVersion
best_oa_location.raw_type	text
best_oa_location.license_id
best_oa_location.is_accepted	False
best_oa_location.is_published	False
best_oa_location.raw_source_name
best_oa_location.landing_page_url	http://arxiv.org/abs/2411.19923
primary_location.id	pmh:oai:arXiv.org:2411.19923
primary_location.is_oa	True
primary_location.source.id	https://openalex.org/S4306400194
primary_location.source.issn
primary_location.source.type	repository
primary_location.source.is_oa	True
primary_location.source.issn_l
primary_location.source.is_core	False
primary_location.source.is_in_doaj	False
primary_location.source.display_name	arXiv (Cornell University)
primary_location.source.host_organization	https://openalex.org/I205783295
primary_location.source.host_organization_name	Cornell University
primary_location.source.host_organization_lineage	https://openalex.org/I205783295
primary_location.license
primary_location.pdf_url	https://arxiv.org/pdf/2411.19923
primary_location.version	submittedVersion
primary_location.raw_type	text
primary_location.license_id
primary_location.is_accepted	False
primary_location.is_published	False
primary_location.raw_source_name
primary_location.landing_page_url	http://arxiv.org/abs/2411.19923
publication_date	2024-11-29
publication_year	2024
referenced_works_count	0
abstract_inverted_index.=	39
abstract_inverted_index.a	81, 132, 139, 145
abstract_inverted_index.\|	37, 41, 43
abstract_inverted_index.We	0
abstract_inverted_index.X)	38
abstract_inverted_index.an	15, 90
abstract_inverted_index.in	32, 60
abstract_inverted_index.is	12, 89, 101
abstract_inverted_index.it	62
abstract_inverted_index.of	4, 72, 126, 134
abstract_inverted_index.on	111, 155
abstract_inverted_index.to	14, 67
abstract_inverted_index.we	130
abstract_inverted_index.(a)	87
abstract_inverted_index.(b)	95
abstract_inverted_index.(c)	99
abstract_inverted_index.(d)	106
abstract_inverted_index.OOD	53, 85
abstract_inverted_index.Our	149
abstract_inverted_index.and	24, 48, 105
abstract_inverted_index.due	13
abstract_inverted_index.for	84, 124
abstract_inverted_index.has	116
abstract_inverted_index.not	64
abstract_inverted_index.set	133
abstract_inverted_index.the	2, 9, 21, 25, 33, 68, 73, 107, 127
abstract_inverted_index.$P(Y	36
abstract_inverted_index.This	28
abstract_inverted_index.\neq	97
abstract_inverted_index.both	20
abstract_inverted_index.does	63
abstract_inverted_index.from	56
abstract_inverted_index.only	144
abstract_inverted_index.task	3
abstract_inverted_index.test	74
abstract_inverted_index.that	61, 137
abstract_inverted_index.work	115
abstract_inverted_index.($X$)	23
abstract_inverted_index.($Z$)	18
abstract_inverted_index.(OOD)	6
abstract_inverted_index.These	78
abstract_inverted_index.While	113
abstract_inverted_index.i.e.,	35
abstract_inverted_index.label	49
abstract_inverted_index.prior	114
abstract_inverted_index.shift	11, 50
abstract_inverted_index.using	143
abstract_inverted_index.where	8
abstract_inverted_index.yield	138
abstract_inverted_index.($Y$).	27
abstract_inverted_index.E_{P(Z	40
abstract_inverted_index.access	66
abstract_inverted_index.assume	65
abstract_inverted_index.create	80
abstract_inverted_index.domain	58
abstract_inverted_index.during	76, 93, 103
abstract_inverted_index.labels	26
abstract_inverted_index.latent	128
abstract_inverted_index.making	45
abstract_inverted_index.simple	141
abstract_inverted_index.single	146
abstract_inverted_index.tasks.	158
abstract_inverted_index.X)}[P(Y	42
abstract_inverted_index.X,Z)]$,	44
abstract_inverted_index.complex	118
abstract_inverted_index.depends	110
abstract_inverted_index.differs	55
abstract_inverted_index.explore	131
abstract_inverted_index.samples	75
abstract_inverted_index.several	156
abstract_inverted_index.approach	150
abstract_inverted_index.consider	1
abstract_inverted_index.multiple	121
abstract_inverted_index.scenario	83
abstract_inverted_index.superior	152
abstract_inverted_index.affecting	19
abstract_inverted_index.benchmark	157
abstract_inverted_index.covariate	47, 69
abstract_inverted_index.developed	117
abstract_inverted_index.empirical	153
abstract_inverted_index.predictor	142
abstract_inverted_index.requiring	120
abstract_inverted_index.training,	94, 104
abstract_inverted_index.training.	77
abstract_inverted_index.variable.	148
abstract_inverted_index.variables	123
abstract_inverted_index.adaptation	59
abstract_inverted_index.additional	122, 147
abstract_inverted_index.conditions	79
abstract_inverted_index.confounder	17, 92
abstract_inverted_index.covariates	22
abstract_inverted_index.introduces	30
abstract_inverted_index.predictive	108
abstract_inverted_index.predictor,	34
abstract_inverted_index.predictors	119
abstract_inverted_index.unobserved	16, 91
abstract_inverted_index.assumptions	51, 136
abstract_inverted_index.challenging	82
abstract_inverted_index.confounding	29
abstract_inverted_index.performance	154
abstract_inverted_index.robustness:	86
abstract_inverted_index.traditional	46, 57
abstract_inverted_index.unavailable	102
abstract_inverted_index.unsuitable.	52
abstract_inverted_index.demonstrates	151
abstract_inverted_index.distribution	10, 70, 109
abstract_inverted_index.surprisingly	140
abstract_inverted_index.$X^\text{te}$	100
abstract_inverted_index.$Z^\text{tr}$	88
abstract_inverted_index.distribution,	129
abstract_inverted_index.heterogeneity	31
abstract_inverted_index.generalization	54
abstract_inverted_index.$P^\text{te}(Z)	96
abstract_inverted_index.($X^\text{te}$)	71
abstract_inverted_index.generalization,	7
abstract_inverted_index.identifiability	125, 135
abstract_inverted_index.P^\text{tr}(Z)$,	98
abstract_inverted_index.$P^\text{te}(Z)$.	112
abstract_inverted_index.out-of-distribution	5
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	4
citation_normalized_percentile