{"id":"https://openalex.org/W4409047785","doi":"https://doi.org/10.1145/3721146.3721961","title":"FlexInfer: Breaking Memory Constraint via Flexible and Efficient Offloading for On-Device LLM Inference","display_name":"FlexInfer: Breaking Memory Constraint via Flexible and Efficient Offloading for On-Device LLM Inference","publication_year":2025,"publication_date":"2025-03-30","ids":{"openalex":"https://openalex.org/W4409047785","doi":"https://doi.org/10.1145/3721146.3721961"},"language":"en","primary_location":{"id":"doi:10.1145/3721146.3721961","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721146.3721961","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3721146.3721961","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 5th Workshop on Machine Learning and Systems","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/10.1145/3721146.3721961","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5091044785","display_name":"Hongchao Du","orcid":"https://orcid.org/0000-0001-7995-1834"},"institutions":[{"id":"https://openalex.org/I168719708","display_name":"City University of Hong Kong","ror":"https://ror.org/03q8dnn23","country_code":"HK","type":"education","lineage":["https://openalex.org/I168719708"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Hongchao Du","raw_affiliation_strings":["City University of Hong Kong, Hong Kong, China"],"affiliations":[{"raw_affiliation_string":"City University of Hong Kong, Hong Kong, China","institution_ids":["https://openalex.org/I168719708"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5061158039","display_name":"Shang-Yu Wu","orcid":"https://orcid.org/0000-0002-1961-143X"},"institutions":[{"id":"https://openalex.org/I168719708","display_name":"City University of Hong Kong","ror":"https://ror.org/03q8dnn23","country_code":"HK","type":"education","lineage":["https://openalex.org/I168719708"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Shangyu Wu","raw_affiliation_strings":["City University of Hong Kong, Hong Kong, China"],"affiliations":[{"raw_affiliation_string":"City University of Hong Kong, Hong Kong, China","institution_ids":["https://openalex.org/I168719708"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025106235","display_name":"Arina Kharlamova","orcid":"https://orcid.org/0009-0001-7620-8567"},"institutions":[{"id":"https://openalex.org/I4210113480","display_name":"Mohamed bin Zayed University of Artificial Intelligence","ror":"https://ror.org/0258gkt32","country_code":"AE","type":"education","lineage":["https://openalex.org/I4210113480"]}],"countries":["AE"],"is_corresponding":false,"raw_author_name":"Arina Kharlamova","raw_affiliation_strings":["Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates"],"affiliations":[{"raw_affiliation_string":"Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates","institution_ids":["https://openalex.org/I4210113480"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5002245169","display_name":"Nan Guan","orcid":"https://orcid.org/0000-0003-3775-911X"},"institutions":[{"id":"https://openalex.org/I168719708","display_name":"City University of Hong Kong","ror":"https://ror.org/03q8dnn23","country_code":"HK","type":"education","lineage":["https://openalex.org/I168719708"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Nan Guan","raw_affiliation_strings":["City University of Hong Kong, Hong Kong, China"],"affiliations":[{"raw_affiliation_string":"City University of Hong Kong, Hong Kong, China","institution_ids":["https://openalex.org/I168719708"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101441768","display_name":"Chun Jason Xue","orcid":"https://orcid.org/0000-0002-6431-9868"},"institutions":[{"id":"https://openalex.org/I4210113480","display_name":"Mohamed bin Zayed University of Artificial Intelligence","ror":"https://ror.org/0258gkt32","country_code":"AE","type":"education","lineage":["https://openalex.org/I4210113480"]}],"countries":["AE"],"is_corresponding":false,"raw_author_name":"Chun Jason Xue","raw_affiliation_strings":["Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates"],"affiliations":[{"raw_affiliation_string":"Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates","institution_ids":["https://openalex.org/I4210113480"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5091044785"],"corresponding_institution_ids":["https://openalex.org/I168719708"],"apc_list":null,"apc_paid":null,"fwci":4.5405,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.9259182,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"56","last_page":"65"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9970999956130981,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9970999956130981,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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":0.996999979019165,"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"}},{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.9947999715805054,"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/computer-science","display_name":"Computer science","score":0.7282482385635376},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.7142112255096436},{"id":"https://openalex.org/keywords/constraint","display_name":"Constraint (computer-aided design)","score":0.6425321698188782},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.33341845870018005},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.279640793800354},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.09246841073036194}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7282482385635376},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7142112255096436},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.6425321698188782},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.33341845870018005},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.279640793800354},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.09246841073036194},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3721146.3721961","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721146.3721961","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3721146.3721961","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 5th Workshop on Machine Learning and Systems","raw_type":"proceedings-article"},{"id":"pmh:oai:pure.atira.dk:publications/2b89eca8-44e4-4ea9-adae-e1c910b1ce02","is_oa":true,"landing_page_url":"https://hdl.handle.net/2031/2b89eca8-44e4-4ea9-adae-e1c910b1ce02","pdf_url":null,"source":{"id":"https://openalex.org/S7407055387","display_name":"CityU Scholars","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Du, H, Wu, S, Kharlamova, A, Guan, N & Xue, C J 2025, FlexInfer : Breaking Memory Constraint via Flexible and Efficient Offloading for On-Device LLM Inference. in EuroMLSys '25 : Proceedings of the 5th Workshop on Machine Learning and Systems. Association for Computing Machinery, pp. 56-65, EuroMLSys '25: 5th Workshop on Machine Learning and Systems, Rotterdam, Netherlands, 30/03/25. https://doi.org/10.1145/3721146.3721961","raw_type":"contributionToPeriodical"}],"best_oa_location":{"id":"doi:10.1145/3721146.3721961","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3721146.3721961","pdf_url":"https://dl.acm.org/doi/pdf/10.1145/3721146.3721961","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 5th Workshop on Machine Learning and Systems","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4409047785.pdf"},"referenced_works_count":16,"referenced_works":["https://openalex.org/W398859631","https://openalex.org/W2757796760","https://openalex.org/W3172198372","https://openalex.org/W4247353671","https://openalex.org/W4321636575","https://openalex.org/W4322718191","https://openalex.org/W4384918448","https://openalex.org/W4385571011","https://openalex.org/W4387321091","https://openalex.org/W4399062099","https://openalex.org/W4401042327","https://openalex.org/W4401306886","https://openalex.org/W4404401018","https://openalex.org/W4404782019","https://openalex.org/W4404782212","https://openalex.org/W4405887239"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"face":[4],"challenges":[5],"for":[6,33],"on-device":[7,34],"inference":[8],"due":[9],"to":[10,16,79],"high":[11,61],"memory":[12,18,45,54],"demands.":[13],"Traditional":[14],"methods":[15,86],"reduce":[17],"usage":[19],"often":[20],"compromise":[21],"performance":[22,62,83],"and":[23,47,56,87],"lack":[24],"adaptability.":[25],"We":[26],"propose":[27],"FlexInfer,":[28],"an":[29],"optimized":[30],"offloading":[31],"framework":[32],"inference,":[35],"addressing":[36],"these":[37],"issues":[38],"with":[39],"techniques":[40],"like":[41],"asynchronous":[42],"prefetching,":[43],"balanced":[44],"locking,":[46],"flexible":[48],"tensor":[49],"preservation.":[50],"These":[51],"strategies":[52],"enhance":[53],"efficiency":[55],"mitigate":[57],"I/O":[58],"bottlenecks,":[59],"ensuring":[60],"within":[63],"user-specified":[64],"resource":[65],"constraints.":[66],"Experiments":[67],"demonstrate":[68],"that":[69],"FlexInfer":[70],"significantly":[71],"improves":[72],"throughput":[73],"under":[74],"limited":[75],"resources,":[76],"achieving":[77],"up":[78],"12.5":[80],"times":[81],"better":[82],"than":[84],"existing":[85],"facilitating":[88],"the":[89],"deployment":[90],"of":[91],"large":[92],"models":[93],"on":[94],"resource-constrained":[95],"devices.":[96]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2026-03-27T05:58:40.876381","created_date":"2025-10-10T00:00:00"}