BAdd: Bias Mitigation through Bias Addition Article Swipe

PDF

Ioannis Sarridis , Christos Koutlis , Symeon Papadopoulos , Christos Diou ·

YOU? · · 2024 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2408.11439

Computer vision (CV) datasets often exhibit biases that are perpetuated by deep learning models. While recent efforts aim to mitigate these biases and foster fair representations, they fail in complex real-world scenarios. In particular, existing methods excel in controlled experiments involving benchmarks with single-attribute injected biases, but struggle with multi-attribute biases being present in well-established CV datasets. Here, we introduce BAdd, a simple yet effective method that allows for learning fair representations invariant to the attributes introducing bias by incorporating features representing these attributes into the backbone. BAdd is evaluated on seven benchmarks and exhibits competitive performance, surpassing state-of-the-art methods on both single- and multi-attribute benchmarks. Notably, BAdd achieves +27.5% and +5.5% absolute accuracy improvements on the challenging multi-attribute benchmarks, FB-Biased-MNIST and CelebA, respectively.

Related Topics

Electrical Engineering

Engineering

Voltage

Concepts

Biasing Electrical engineering Engineering Voltage

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2408.11439
PDF: https://arxiv.org/pdf/2408.11439
OA Status: green
Related Works: 10
OpenAlex ID: https://openalex.org/W4405426124

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4405426124

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2408.11439

Digital Object Identifier
Title: BAdd: Bias Mitigation through Bias Addition

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2024

Year of publication
Publication date: 2024-08-21

Full publication date if available
Authors: Ioannis Sarridis, Christos Koutlis, Symeon Papadopoulos, Christos Diou

List of authors in order
Landing page: https://arxiv.org/abs/2408.11439

Publisher landing page
PDF URL: https://arxiv.org/pdf/2408.11439

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/2408.11439

Direct OA link when available
Concepts: Biasing, Electrical engineering, Engineering, Voltage

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/W4405426124
doi	https://doi.org/10.48550/arxiv.2408.11439
ids.doi	https://doi.org/10.48550/arxiv.2408.11439
ids.openalex	https://openalex.org/W4405426124
fwci
type	preprint
title	BAdd: Bias Mitigation through Bias Addition
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
topics[0].id	https://openalex.org/T11689
topics[0].field.id	https://openalex.org/fields/17
topics[0].field.display_name	Computer Science
topics[0].score	0.444599986076355
topics[0].domain.id	https://openalex.org/domains/3
topics[0].domain.display_name	Physical Sciences
topics[0].subfield.id	https://openalex.org/subfields/1702
topics[0].subfield.display_name	Artificial Intelligence
topics[0].display_name	Adversarial Robustness in Machine Learning
is_xpac	False
apc_list
apc_paid
concepts[0].id	https://openalex.org/C20254490
concepts[0].level	3
concepts[0].score	0.3293790817260742
concepts[0].wikidata	https://www.wikidata.org/wiki/Q719550
concepts[0].display_name	Biasing
concepts[1].id	https://openalex.org/C119599485
concepts[1].level	1
concepts[1].score	0.10957059264183044
concepts[1].wikidata	https://www.wikidata.org/wiki/Q43035
concepts[1].display_name	Electrical engineering
concepts[2].id	https://openalex.org/C127413603
concepts[2].level	0
concepts[2].score	0.10397112369537354
concepts[2].wikidata	https://www.wikidata.org/wiki/Q11023
concepts[2].display_name	Engineering
concepts[3].id	https://openalex.org/C165801399
concepts[3].level	2
concepts[3].score	0.0
concepts[3].wikidata	https://www.wikidata.org/wiki/Q25428
concepts[3].display_name	Voltage
keywords[0].id	https://openalex.org/keywords/biasing
keywords[0].score	0.3293790817260742
keywords[0].display_name	Biasing
keywords[1].id	https://openalex.org/keywords/electrical-engineering
keywords[1].score	0.10957059264183044
keywords[1].display_name	Electrical engineering
keywords[2].id	https://openalex.org/keywords/engineering
keywords[2].score	0.10397112369537354
keywords[2].display_name	Engineering
language	en
locations[0].id	pmh:oai:arXiv.org:2408.11439
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/2408.11439
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/2408.11439
locations[1].id	doi:10.48550/arxiv.2408.11439
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.2408.11439
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5005706607
authorships[0].author.orcid
authorships[0].author.display_name	Ioannis Sarridis
authorships[0].author_position	first
authorships[0].raw_author_name	Sarridis, Ioannis
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5024874767
authorships[1].author.orcid	https://orcid.org/0000-0003-3682-408X
authorships[1].author.display_name	Christos Koutlis
authorships[1].author_position	middle
authorships[1].raw_author_name	Koutlis, Christos
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5013616365
authorships[2].author.orcid	https://orcid.org/0000-0002-5441-7341
authorships[2].author.display_name	Symeon Papadopoulos
authorships[2].author_position	middle
authorships[2].raw_author_name	Papadopoulos, Symeon
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5004953619
authorships[3].author.orcid	https://orcid.org/0000-0002-2461-1928
authorships[3].author.display_name	Christos Diou
authorships[3].author_position	last
authorships[3].raw_author_name	Diou, Christos
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/2408.11439
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-10-10T00:00:00
display_name	BAdd: Bias Mitigation through Bias Addition
has_fulltext	False
is_retracted	False
updated_date	2025-11-06T06:51:31.235846
primary_topic.id	https://openalex.org/T11689
primary_topic.field.id	https://openalex.org/fields/17
primary_topic.field.display_name	Computer Science
primary_topic.score	0.444599986076355
primary_topic.domain.id	https://openalex.org/domains/3
primary_topic.domain.display_name	Physical Sciences
primary_topic.subfield.id	https://openalex.org/subfields/1702
primary_topic.subfield.display_name	Artificial Intelligence
primary_topic.display_name	Adversarial Robustness in Machine Learning
related_works	https://openalex.org/W2069427488, https://openalex.org/W4281694563, https://openalex.org/W2080696413, https://openalex.org/W1506140395, https://openalex.org/W4232799642, https://openalex.org/W2912082923, https://openalex.org/W2744827311, https://openalex.org/W2558598037, https://openalex.org/W2163182355, https://openalex.org/W2376107070
cited_by_count	0
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2408.11439
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/2408.11439
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/2408.11439
primary_location.id	pmh:oai:arXiv.org:2408.11439
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/2408.11439
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/2408.11439
publication_date	2024-08-21
publication_year	2024
referenced_works_count	0
abstract_inverted_index.a	61
abstract_inverted_index.CV	55
abstract_inverted_index.In	32
abstract_inverted_index.by	10, 78
abstract_inverted_index.in	28, 37, 53
abstract_inverted_index.is	88
abstract_inverted_index.on	90, 100, 115
abstract_inverted_index.to	18, 73
abstract_inverted_index.we	58
abstract_inverted_index.aim	17
abstract_inverted_index.and	22, 93, 103, 110, 121
abstract_inverted_index.are	8
abstract_inverted_index.but	46
abstract_inverted_index.for	68
abstract_inverted_index.the	74, 85, 116
abstract_inverted_index.yet	63
abstract_inverted_index.(CV)	2
abstract_inverted_index.BAdd	87, 107
abstract_inverted_index.bias	77
abstract_inverted_index.both	101
abstract_inverted_index.deep	11
abstract_inverted_index.fail	27
abstract_inverted_index.fair	24, 70
abstract_inverted_index.into	84
abstract_inverted_index.that	7, 66
abstract_inverted_index.they	26
abstract_inverted_index.with	42, 48
abstract_inverted_index.+5.5%	111
abstract_inverted_index.BAdd,	60
abstract_inverted_index.Here,	57
abstract_inverted_index.While	14
abstract_inverted_index.being	51
abstract_inverted_index.excel	36
abstract_inverted_index.often	4
abstract_inverted_index.seven	91
abstract_inverted_index.these	20, 82
abstract_inverted_index.+27.5%	109
abstract_inverted_index.allows	67
abstract_inverted_index.biases	6, 21, 50
abstract_inverted_index.foster	23
abstract_inverted_index.method	65
abstract_inverted_index.recent	15
abstract_inverted_index.simple	62
abstract_inverted_index.vision	1
abstract_inverted_index.CelebA,	122
abstract_inverted_index.biases,	45
abstract_inverted_index.complex	29
abstract_inverted_index.efforts	16
abstract_inverted_index.exhibit	5
abstract_inverted_index.methods	35, 99
abstract_inverted_index.models.	13
abstract_inverted_index.present	52
abstract_inverted_index.single-	102
abstract_inverted_index.Computer	0
abstract_inverted_index.Notably,	106
abstract_inverted_index.absolute	112
abstract_inverted_index.accuracy	113
abstract_inverted_index.achieves	108
abstract_inverted_index.datasets	3
abstract_inverted_index.exhibits	94
abstract_inverted_index.existing	34
abstract_inverted_index.features	80
abstract_inverted_index.injected	44
abstract_inverted_index.learning	12, 69
abstract_inverted_index.mitigate	19
abstract_inverted_index.struggle	47
abstract_inverted_index.backbone.	86
abstract_inverted_index.datasets.	56
abstract_inverted_index.effective	64
abstract_inverted_index.evaluated	89
abstract_inverted_index.introduce	59
abstract_inverted_index.invariant	72
abstract_inverted_index.involving	40
abstract_inverted_index.attributes	75, 83
abstract_inverted_index.benchmarks	41, 92
abstract_inverted_index.controlled	38
abstract_inverted_index.real-world	30
abstract_inverted_index.scenarios.	31
abstract_inverted_index.surpassing	97
abstract_inverted_index.benchmarks,	119
abstract_inverted_index.benchmarks.	105
abstract_inverted_index.challenging	117
abstract_inverted_index.competitive	95
abstract_inverted_index.experiments	39
abstract_inverted_index.introducing	76
abstract_inverted_index.particular,	33
abstract_inverted_index.perpetuated	9
abstract_inverted_index.improvements	114
abstract_inverted_index.performance,	96
abstract_inverted_index.representing	81
abstract_inverted_index.incorporating	79
abstract_inverted_index.respectively.	123
abstract_inverted_index.FB-Biased-MNIST	120
abstract_inverted_index.multi-attribute	49, 104, 118
abstract_inverted_index.representations	71
abstract_inverted_index.representations,	25
abstract_inverted_index.single-attribute	43
abstract_inverted_index.state-of-the-art	98
abstract_inverted_index.well-established	54
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	4
citation_normalized_percentile