Resilient LLM-Empowered Semantic MAC Protocols via Zero-Shot Adaptation and Knowledge Distillation Article Swipe

PDF

Yong-Jun Kim , Jihong Park , Mehdi Bennis , Junil Choi ·

YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2505.21518

Neural network-based medium access control (MAC) protocol models (NPMs) improve goodput through site-specific operations but are vulnerable to shifts from their training network environments, such as changes in the number of user equipments (UEs) severely degrade goodput. To enhance resilience against such environmental shifts, we propose three novel semantic MAC protocol frameworks empowered by large language models (LLMs). First, we introduce a token-based protocol model (TPM), where an LLM generates MAC signaling messages. By editing LLM instruction prompts, TPM enables instant adaptation, which can be further enhanced by TextGrad, an LLM-based automated prompt optimizer. TPM inference is fast but coarse due to the lack of real interactions with the changed environment, and computationally intensive due to the large size of the LLM. To improve goodput and computation efficiency, we develop T2NPM, which transfers and augments TPM knowledge into an NPM via knowledge distillation (KD). Integrating TPM and T2NPM, we propose T3NPM, which employs TPM in the early phase and switches to T2NPM later. To optimize this phase switching, we design a novel metric of meta-resilience, which quantifies resilience to unknown target goodput after environmental shifts. Simulations corroborate that T3NPM achieves 20.56% higher meta-resilience than NPM with 19.8x lower computation cost than TPM in FLOPS.

Related Topics

Concepts

No concepts available.

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2505.21518
PDF: https://arxiv.org/pdf/2505.21518
OA Status: green
OpenAlex ID: https://openalex.org/W4416045976

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4416045976

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2505.21518

Digital Object Identifier
Title: Resilient LLM-Empowered Semantic MAC Protocols via Zero-Shot Adaptation and Knowledge Distillation

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2025

Year of publication
Publication date: 2025-05-22

Full publication date if available
Authors: Yong-Jun Kim, Jihong Park, Mehdi Bennis, Junil Choi

List of authors in order
Landing page: https://arxiv.org/abs/2505.21518

Publisher landing page
PDF URL: https://arxiv.org/pdf/2505.21518

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/2505.21518

Direct OA link when available
Cited by: 0

Total citation count in OpenAlex

Full payload

id	https://openalex.org/W4416045976
doi	https://doi.org/10.48550/arxiv.2505.21518
ids.doi	https://doi.org/10.48550/arxiv.2505.21518
ids.openalex	https://openalex.org/W4416045976
fwci
type	preprint
title	Resilient LLM-Empowered Semantic MAC Protocols via Zero-Shot Adaptation and Knowledge Distillation
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:2505.21518
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/2505.21518
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/2505.21518
locations[1].id	doi:10.48550/arxiv.2505.21518
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.2505.21518
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5101852904
authorships[0].author.orcid	https://orcid.org/0000-0002-8472-7770
authorships[0].author.display_name	Yong-Jun Kim
authorships[0].author_position	first
authorships[0].raw_author_name	Kim, Yongjun
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5027907258
authorships[1].author.orcid	https://orcid.org/0000-0001-7623-6552
authorships[1].author.display_name	Jihong Park
authorships[1].author_position	middle
authorships[1].raw_author_name	Park, Jihong
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5061429095
authorships[2].author.orcid	https://orcid.org/0000-0003-0261-0171
authorships[2].author.display_name	Mehdi Bennis
authorships[2].author_position	middle
authorships[2].raw_author_name	Bennis, Mehdi
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5065248740
authorships[3].author.orcid	https://orcid.org/0000-0002-9862-9020
authorships[3].author.display_name	Junil Choi
authorships[3].author_position	last
authorships[3].raw_author_name	Choi, Junil
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/2505.21518
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-10-10T00:00:00
display_name	Resilient LLM-Empowered Semantic MAC Protocols via Zero-Shot Adaptation and Knowledge Distillation
has_fulltext	False
is_retracted	False
updated_date	2025-11-28T11:01:35.901026
primary_topic
cited_by_count	0
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2505.21518
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/2505.21518
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/2505.21518
primary_location.id	pmh:oai:arXiv.org:2505.21518
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/2505.21518
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/2505.21518
publication_date	2025-05-22
publication_year	2025
referenced_works_count	0
abstract_inverted_index.a	61, 170
abstract_inverted_index.By	73
abstract_inverted_index.To	37, 122, 163
abstract_inverted_index.an	67, 89, 138
abstract_inverted_index.as	25
abstract_inverted_index.be	84
abstract_inverted_index.by	53, 87
abstract_inverted_index.in	27, 154, 202
abstract_inverted_index.is	96
abstract_inverted_index.of	30, 104, 119, 173
abstract_inverted_index.to	17, 101, 115, 160, 178
abstract_inverted_index.we	44, 59, 128, 148, 168
abstract_inverted_index.LLM	68, 75
abstract_inverted_index.MAC	49, 70
abstract_inverted_index.NPM	139, 194
abstract_inverted_index.TPM	78, 94, 135, 145, 153, 201
abstract_inverted_index.and	111, 125, 133, 146, 158
abstract_inverted_index.are	15
abstract_inverted_index.but	14, 98
abstract_inverted_index.can	83
abstract_inverted_index.due	100, 114
abstract_inverted_index.the	28, 102, 108, 116, 120, 155
abstract_inverted_index.via	140
abstract_inverted_index.LLM.	121
abstract_inverted_index.cost	199
abstract_inverted_index.fast	97
abstract_inverted_index.from	19
abstract_inverted_index.into	137
abstract_inverted_index.lack	103
abstract_inverted_index.real	105
abstract_inverted_index.size	118
abstract_inverted_index.such	24, 41
abstract_inverted_index.than	193, 200
abstract_inverted_index.that	187
abstract_inverted_index.this	165
abstract_inverted_index.user	31
abstract_inverted_index.with	107, 195
abstract_inverted_index.(KD).	143
abstract_inverted_index.(MAC)	5
abstract_inverted_index.(UEs)	33
abstract_inverted_index.19.8x	196
abstract_inverted_index.T2NPM	161
abstract_inverted_index.T3NPM	188
abstract_inverted_index.after	182
abstract_inverted_index.early	156
abstract_inverted_index.large	54, 117
abstract_inverted_index.lower	197
abstract_inverted_index.model	64
abstract_inverted_index.novel	47, 171
abstract_inverted_index.phase	157, 166
abstract_inverted_index.their	20
abstract_inverted_index.three	46
abstract_inverted_index.where	66
abstract_inverted_index.which	82, 131, 151, 175
abstract_inverted_index.(NPMs)	8
abstract_inverted_index.(TPM),	65
abstract_inverted_index.20.56%	190
abstract_inverted_index.FLOPS.	203
abstract_inverted_index.First,	58
abstract_inverted_index.Neural	0
abstract_inverted_index.T2NPM,	130, 147
abstract_inverted_index.T3NPM,	150
abstract_inverted_index.access	3
abstract_inverted_index.coarse	99
abstract_inverted_index.design	169
abstract_inverted_index.higher	191
abstract_inverted_index.later.	162
abstract_inverted_index.medium	2
abstract_inverted_index.metric	172
abstract_inverted_index.models	7, 56
abstract_inverted_index.number	29
abstract_inverted_index.prompt	92
abstract_inverted_index.shifts	18
abstract_inverted_index.target	180
abstract_inverted_index.(LLMs).	57
abstract_inverted_index.against	40
abstract_inverted_index.changed	109
abstract_inverted_index.changes	26
abstract_inverted_index.control	4
abstract_inverted_index.degrade	35
abstract_inverted_index.develop	129
abstract_inverted_index.editing	74
abstract_inverted_index.employs	152
abstract_inverted_index.enables	79
abstract_inverted_index.enhance	38
abstract_inverted_index.further	85
abstract_inverted_index.goodput	10, 124, 181
abstract_inverted_index.improve	9, 123
abstract_inverted_index.instant	80
abstract_inverted_index.network	22
abstract_inverted_index.propose	45, 149
abstract_inverted_index.shifts,	43
abstract_inverted_index.shifts.	184
abstract_inverted_index.through	11
abstract_inverted_index.unknown	179
abstract_inverted_index.achieves	189
abstract_inverted_index.augments	134
abstract_inverted_index.enhanced	86
abstract_inverted_index.goodput.	36
abstract_inverted_index.language	55
abstract_inverted_index.optimize	164
abstract_inverted_index.prompts,	77
abstract_inverted_index.protocol	6, 50, 63
abstract_inverted_index.semantic	48
abstract_inverted_index.severely	34
abstract_inverted_index.switches	159
abstract_inverted_index.training	21
abstract_inverted_index.LLM-based	90
abstract_inverted_index.TextGrad,	88
abstract_inverted_index.automated	91
abstract_inverted_index.empowered	52
abstract_inverted_index.generates	69
abstract_inverted_index.inference	95
abstract_inverted_index.intensive	113
abstract_inverted_index.introduce	60
abstract_inverted_index.knowledge	136, 141
abstract_inverted_index.messages.	72
abstract_inverted_index.signaling	71
abstract_inverted_index.transfers	132
abstract_inverted_index.equipments	32
abstract_inverted_index.frameworks	51
abstract_inverted_index.operations	13
abstract_inverted_index.optimizer.	93
abstract_inverted_index.quantifies	176
abstract_inverted_index.resilience	39, 177
abstract_inverted_index.switching,	167
abstract_inverted_index.vulnerable	16
abstract_inverted_index.Integrating	144
abstract_inverted_index.Simulations	185
abstract_inverted_index.adaptation,	81
abstract_inverted_index.computation	126, 198
abstract_inverted_index.corroborate	186
abstract_inverted_index.efficiency,	127
abstract_inverted_index.instruction	76
abstract_inverted_index.token-based	62
abstract_inverted_index.distillation	142
abstract_inverted_index.environment,	110
abstract_inverted_index.interactions	106
abstract_inverted_index.environmental	42, 183
abstract_inverted_index.environments,	23
abstract_inverted_index.network-based	1
abstract_inverted_index.site-specific	12
abstract_inverted_index.computationally	112
abstract_inverted_index.meta-resilience	192
abstract_inverted_index.meta-resilience,	174
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	4
citation_normalized_percentile