{"id":"https://openalex.org/W4416920761","doi":"https://doi.org/10.1145/3737899.3768528","title":"FedHO: Memory-Efficient Federated Fine-Tuning for Large Models via Hybrid Gradient Computation","display_name":"FedHO: Memory-Efficient Federated Fine-Tuning for Large Models via Hybrid Gradient Computation","publication_year":2025,"publication_date":"2025-11-04","ids":{"openalex":"https://openalex.org/W4416920761","doi":"https://doi.org/10.1145/3737899.3768528"},"language":null,"primary_location":{"id":"doi:10.1145/3737899.3768528","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3737899.3768528","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3737899.3768528","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Federated Learning and Edge AI for Privacy and Mobility","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://dl.acm.org/doi/pdf/10.1145/3737899.3768528","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5113467245","display_name":"Yinan Zhang","orcid":"https://orcid.org/0009-0004-5943-5647"},"institutions":[{"id":"https://openalex.org/I14243506","display_name":"Hong Kong Polytechnic University","ror":"https://ror.org/0030zas98","country_code":"HK","type":"education","lineage":["https://openalex.org/I14243506"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Yinan Zhang","raw_affiliation_strings":["The Hong Kong Polytechnic University Hong Kong, Hong Kong"],"affiliations":[{"raw_affiliation_string":"The Hong Kong Polytechnic University Hong Kong, Hong Kong","institution_ids":["https://openalex.org/I14243506"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100740023","display_name":"Jiannong Cao","orcid":"https://orcid.org/0000-0002-2725-2529"},"institutions":[{"id":"https://openalex.org/I14243506","display_name":"Hong Kong Polytechnic University","ror":"https://ror.org/0030zas98","country_code":"HK","type":"education","lineage":["https://openalex.org/I14243506"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Jiannong Cao","raw_affiliation_strings":["The Hong Kong Polytechnic University Hong Kong, Hong Kong"],"affiliations":[{"raw_affiliation_string":"The Hong Kong Polytechnic University Hong Kong, Hong Kong","institution_ids":["https://openalex.org/I14243506"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5106557838","display_name":"Mingjin Zhang","orcid":"https://orcid.org/0000-0002-1653-108X"},"institutions":[{"id":"https://openalex.org/I14243506","display_name":"Hong Kong Polytechnic University","ror":"https://ror.org/0030zas98","country_code":"HK","type":"education","lineage":["https://openalex.org/I14243506"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Mingjin Zhang","raw_affiliation_strings":["The Hong Kong Polytechnic University Hong Kong, Hong Kong"],"affiliations":[{"raw_affiliation_string":"The Hong Kong Polytechnic University Hong Kong, Hong Kong","institution_ids":["https://openalex.org/I14243506"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5081781438","display_name":"Ruosong Yang","orcid":"https://orcid.org/0000-0002-9483-4000"},"institutions":[{"id":"https://openalex.org/I146617529","display_name":"Applied Science and Technology Research Institute","ror":"https://ror.org/03xmkea05","country_code":"CN","type":"facility","lineage":["https://openalex.org/I146617529"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ruosong Yang","raw_affiliation_strings":["China Mobile (Hong Kong) Innovation and Research Institute Hong Kong, Hong Kong"],"affiliations":[{"raw_affiliation_string":"China Mobile (Hong Kong) Innovation and Research Institute Hong Kong, Hong Kong","institution_ids":["https://openalex.org/I146617529"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5113467245"],"corresponding_institution_ids":["https://openalex.org/I14243506"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.20254602,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"85","last_page":"92"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.2529999911785126,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.2529999911785126,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.1177000030875206,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10273","display_name":"IoT and Edge/Fog Computing","score":0.09989999979734421,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.7649999856948853},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.6855000257492065},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.6258999705314636},{"id":"https://openalex.org/keywords/convergence","display_name":"Convergence (economics)","score":0.5769000053405762},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.5424000024795532},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.5360999703407288},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.5234000086784363}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8149999976158142},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.7649999856948853},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.6855000257492065},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.6258999705314636},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.5769000053405762},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.5424000024795532},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.5360999703407288},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.5234000086784363},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.5217999815940857},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.4253000020980835},{"id":"https://openalex.org/C155032097","wikidata":"https://www.wikidata.org/wiki/Q798503","display_name":"Backpropagation","level":3,"score":0.4115000069141388},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.3887999951839447},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.3156000077724457},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.2865000069141388},{"id":"https://openalex.org/C189950617","wikidata":"https://www.wikidata.org/wiki/Q937228","display_name":"Property (philosophy)","level":2,"score":0.2799000144004822},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.27720001339912415},{"id":"https://openalex.org/C115680565","wikidata":"https://www.wikidata.org/wiki/Q5977448","display_name":"Gradient method","level":2,"score":0.26269999146461487},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2615000009536743},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.25859999656677246}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3737899.3768528","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3737899.3768528","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3737899.3768528","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Federated Learning and Edge AI for Privacy and Mobility","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3737899.3768528","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3737899.3768528","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3737899.3768528","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Federated Learning and Edge AI for Privacy and Mobility","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320322598","display_name":"Hong Kong Polytechnic University","ror":"https://ror.org/0030zas98"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416920761.pdf","grobid_xml":"https://content.openalex.org/works/W4416920761.grobid-xml"},"referenced_works_count":5,"referenced_works":["https://openalex.org/W2965862774","https://openalex.org/W3133814152","https://openalex.org/W4226479888","https://openalex.org/W4401863415","https://openalex.org/W4405003490"],"related_works":[],"abstract_inverted_index":{"Federated":[0],"fine-tuning":[1],"of":[2],"Large":[3],"Language":[4],"Models":[5],"(LLMs)":[6],"is":[7],"crucial":[8],"for":[9,23,71],"Edge":[10],"AI,":[11],"yet":[12],"the":[13,89,96],"immense":[14],"memory":[15,36,108,127],"required":[16],"by":[17,38],"standard":[18],"backpropagation":[19],"creates":[20],"a":[21,63,102],"bottleneck":[22],"resource-constrained":[24],"edge":[25],"devices.":[26],"While":[27],"recent":[28],"efforts":[29],"adopt":[30],"zerothorder":[31],"optimization":[32,69,93],"(ZOO)":[33],"to":[34,49,130],"alleviate":[35],"overhead":[37,128],"eliminating":[39],"gradient":[40,72],"computation,":[41],"simply":[42],"applying":[43],"ZOO":[44,120],"in":[45,74],"FedLLM":[46],"may":[47],"lead":[48],"catastrophic":[50],"training":[51],"instability":[52],"(e.g.,":[53],"loss":[54],"divergence).":[55],"To":[56],"address":[57],"this":[58],"issue,":[59],"we":[60],"propose":[61],"FedHO,":[62],"novel":[64],"framework":[65],"that":[66,113],"introduces":[67],"hybrid":[68],"(HO)":[70],"computation":[73],"federated":[75],"LLM":[76],"fine-tuning.":[77],"HO":[78],"dynamically":[79],"groups":[80],"model":[81,105],"parameters":[82],"into":[83],"high-":[84],"and":[85,95,107,122,125],"low-importance":[86],"groups,":[87],"updating":[88],"former":[90],"via":[91,98],"first-order":[92],"(FOO)":[94],"latter":[97],"memory-efficient":[99],"ZOO,":[100],"striking":[101],"balance":[103],"between":[104],"performance":[106],"overhead.":[109],"Extensive":[110],"experiments":[111],"indicate":[112],"FedHO":[114],"offers":[115],"better":[116],"stability":[117],"than":[118],"state-of-the-art":[119],"solutions,":[121],"faster":[123],"convergence":[124],"lower":[126],"compared":[129],"BP-based":[131],"FedAvg.":[132]},"counts_by_year":[],"updated_date":"2026-03-27T05:58:40.876381","created_date":"2025-12-02T00:00:00"}