SCU: An Efficient Machine Unlearning Scheme for Deep Learning Enabled Semantic Communications Article Swipe

PDF

Weiqi Wang , Zhiyi Tian , Chenhan Zhang , Shui Yu ·

YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2502.19785

Deep learning (DL) enabled semantic communications leverage DL to train encoders and decoders (codecs) to extract and recover semantic information. However, most semantic training datasets contain personal private information. Such concerns call for enormous requirements for specified data erasure from semantic codecs when previous users hope to move their data from the semantic system. {Existing machine unlearning solutions remove data contribution from trained models, yet usually in supervised sole model scenarios. These methods are infeasible in semantic communications that often need to jointly train unsupervised encoders and decoders.} In this paper, we investigate the unlearning problem in DL-enabled semantic communications and propose a semantic communication unlearning (SCU) scheme to tackle the problem. {SCU includes two key components. Firstly,} we customize the joint unlearning method for semantic codecs, including the encoder and decoder, by minimizing mutual information between the learned semantic representation and the erased samples. {Secondly,} to compensate for semantic model utility degradation caused by unlearning, we propose a contrastive compensation method, which considers the erased data as the negative samples and the remaining data as the positive samples to retrain the unlearned semantic models contrastively. Theoretical analysis and extensive experimental results on three representative datasets demonstrate the effectiveness and efficiency of our proposed methods.

Related Topics

Enigma Machine

Machine Learning

The Myth Of The Machine

The Second Machine Age

Florence And The Machine

The Emotion Machine

M249 Light Machine Gun

Rube Goldberg Machine

Concepts

No concepts available.

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2502.19785
PDF: https://arxiv.org/pdf/2502.19785
OA Status: green
OpenAlex ID: https://openalex.org/W4416149819

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4416149819

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2502.19785

Digital Object Identifier
Title: SCU: An Efficient Machine Unlearning Scheme for Deep Learning Enabled Semantic Communications

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2025

Year of publication
Publication date: 2025-02-27

Full publication date if available
Authors: Weiqi Wang, Zhiyi Tian, Chenhan Zhang, Shui Yu

List of authors in order
Landing page: https://arxiv.org/abs/2502.19785

Publisher landing page
PDF URL: https://arxiv.org/pdf/2502.19785

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/2502.19785

Direct OA link when available
Cited by: 0

Total citation count in OpenAlex

Full payload

id	https://openalex.org/W4416149819
doi	https://doi.org/10.48550/arxiv.2502.19785
ids.doi	https://doi.org/10.48550/arxiv.2502.19785
ids.openalex	https://openalex.org/W4416149819
fwci
type	preprint
title	SCU: An Efficient Machine Unlearning Scheme for Deep Learning Enabled Semantic Communications
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
is_xpac	False
apc_list
apc_paid
language	en
locations[0].id	pmh:oai:arXiv.org:2502.19785
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/2502.19785
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/2502.19785
locations[1].id	doi:10.48550/arxiv.2502.19785
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	cc-by
locations[1].pdf_url
locations[1].version
locations[1].raw_type	article
locations[1].license_id	https://openalex.org/licenses/cc-by
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.2502.19785
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5101907050
authorships[0].author.orcid	https://orcid.org/0000-0002-7905-3126
authorships[0].author.display_name	Weiqi Wang
authorships[0].author_position	first
authorships[0].raw_author_name	Wang, Weiqi
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5059567156
authorships[1].author.orcid	https://orcid.org/0000-0001-8905-0941
authorships[1].author.display_name	Zhiyi Tian
authorships[1].author_position	middle
authorships[1].raw_author_name	Tian, Zhiyi
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5061304518
authorships[2].author.orcid	https://orcid.org/0000-0002-2352-0485
authorships[2].author.display_name	Chenhan Zhang
authorships[2].author_position	middle
authorships[2].raw_author_name	Zhang, Chenhan
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5005228053
authorships[3].author.orcid	https://orcid.org/0000-0003-4485-6743
authorships[3].author.display_name	Shui Yu
authorships[3].author_position	last
authorships[3].raw_author_name	Yu, Shui
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/2502.19785
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-10-10T00:00:00
display_name	SCU: An Efficient Machine Unlearning Scheme for Deep Learning Enabled Semantic Communications
has_fulltext	False
is_retracted	False
updated_date	2025-11-28T06:25:34.151263
primary_topic
cited_by_count	0
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2502.19785
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/2502.19785
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/2502.19785
primary_location.id	pmh:oai:arXiv.org:2502.19785
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/2502.19785
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/2502.19785
publication_date	2025-02-27
publication_year	2025
referenced_works_count	0
abstract_inverted_index.a	102, 158
abstract_inverted_index.DL	7
abstract_inverted_index.In	88
abstract_inverted_index.as	167, 175
abstract_inverted_index.by	132, 154
abstract_inverted_index.in	66, 75, 96
abstract_inverted_index.of	201
abstract_inverted_index.on	192
abstract_inverted_index.to	8, 14, 46, 81, 108, 146, 179
abstract_inverted_index.we	91, 118, 156
abstract_inverted_index.and	11, 16, 86, 100, 130, 141, 171, 188, 199
abstract_inverted_index.are	73
abstract_inverted_index.for	32, 35, 124, 148
abstract_inverted_index.key	115
abstract_inverted_index.our	202
abstract_inverted_index.the	51, 93, 110, 120, 128, 137, 142, 164, 168, 172, 176, 181, 197
abstract_inverted_index.two	114
abstract_inverted_index.yet	64
abstract_inverted_index.(DL)	2
abstract_inverted_index.Deep	0
abstract_inverted_index.Such	29
abstract_inverted_index.call	31
abstract_inverted_index.data	37, 49, 59, 166, 174
abstract_inverted_index.from	39, 50, 61
abstract_inverted_index.hope	45
abstract_inverted_index.most	21
abstract_inverted_index.move	47
abstract_inverted_index.need	80
abstract_inverted_index.sole	68
abstract_inverted_index.that	78
abstract_inverted_index.this	89
abstract_inverted_index.when	42
abstract_inverted_index.{SCU	112
abstract_inverted_index.(SCU)	106
abstract_inverted_index.These	71
abstract_inverted_index.joint	121
abstract_inverted_index.model	69, 150
abstract_inverted_index.often	79
abstract_inverted_index.their	48
abstract_inverted_index.three	193
abstract_inverted_index.train	9, 83
abstract_inverted_index.users	44
abstract_inverted_index.which	162
abstract_inverted_index.caused	153
abstract_inverted_index.codecs	41
abstract_inverted_index.erased	143, 165
abstract_inverted_index.method	123
abstract_inverted_index.models	184
abstract_inverted_index.mutual	134
abstract_inverted_index.paper,	90
abstract_inverted_index.remove	58
abstract_inverted_index.scheme	107
abstract_inverted_index.tackle	109
abstract_inverted_index.between	136
abstract_inverted_index.codecs,	126
abstract_inverted_index.contain	25
abstract_inverted_index.enabled	3
abstract_inverted_index.encoder	129
abstract_inverted_index.erasure	38
abstract_inverted_index.extract	15
abstract_inverted_index.jointly	82
abstract_inverted_index.learned	138
abstract_inverted_index.machine	55
abstract_inverted_index.method,	161
abstract_inverted_index.methods	72
abstract_inverted_index.models,	63
abstract_inverted_index.private	27
abstract_inverted_index.problem	95
abstract_inverted_index.propose	101, 157
abstract_inverted_index.recover	17
abstract_inverted_index.results	191
abstract_inverted_index.retrain	180
abstract_inverted_index.samples	170, 178
abstract_inverted_index.system.	53
abstract_inverted_index.trained	62
abstract_inverted_index.usually	65
abstract_inverted_index.utility	151
abstract_inverted_index.(codecs)	13
abstract_inverted_index.However,	20
abstract_inverted_index.analysis	187
abstract_inverted_index.concerns	30
abstract_inverted_index.datasets	24, 195
abstract_inverted_index.decoder,	131
abstract_inverted_index.decoders	12
abstract_inverted_index.encoders	10, 85
abstract_inverted_index.enormous	33
abstract_inverted_index.includes	113
abstract_inverted_index.learning	1
abstract_inverted_index.leverage	6
abstract_inverted_index.methods.	204
abstract_inverted_index.negative	169
abstract_inverted_index.personal	26
abstract_inverted_index.positive	177
abstract_inverted_index.previous	43
abstract_inverted_index.problem.	111
abstract_inverted_index.proposed	203
abstract_inverted_index.samples.	144
abstract_inverted_index.semantic	4, 18, 22, 40, 52, 76, 98, 103, 125, 139, 149, 183
abstract_inverted_index.training	23
abstract_inverted_index.Firstly,}	117
abstract_inverted_index.considers	163
abstract_inverted_index.customize	119
abstract_inverted_index.extensive	189
abstract_inverted_index.including	127
abstract_inverted_index.remaining	173
abstract_inverted_index.solutions	57
abstract_inverted_index.specified	36
abstract_inverted_index.unlearned	182
abstract_inverted_index.{Existing	54
abstract_inverted_index.DL-enabled	97
abstract_inverted_index.compensate	147
abstract_inverted_index.decoders.}	87
abstract_inverted_index.efficiency	200
abstract_inverted_index.infeasible	74
abstract_inverted_index.minimizing	133
abstract_inverted_index.scenarios.	70
abstract_inverted_index.supervised	67
abstract_inverted_index.unlearning	56, 94, 105, 122
abstract_inverted_index.Theoretical	186
abstract_inverted_index.components.	116
abstract_inverted_index.contrastive	159
abstract_inverted_index.degradation	152
abstract_inverted_index.demonstrate	196
abstract_inverted_index.information	135
abstract_inverted_index.investigate	92
abstract_inverted_index.unlearning,	155
abstract_inverted_index.{Secondly,}	145
abstract_inverted_index.compensation	160
abstract_inverted_index.contribution	60
abstract_inverted_index.experimental	190
abstract_inverted_index.information.	19, 28
abstract_inverted_index.requirements	34
abstract_inverted_index.unsupervised	84
abstract_inverted_index.communication	104
abstract_inverted_index.effectiveness	198
abstract_inverted_index.communications	5, 77, 99
abstract_inverted_index.contrastively.	185
abstract_inverted_index.representation	140
abstract_inverted_index.representative	194
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	4
citation_normalized_percentile