{"id":"https://openalex.org/W7161115507","doi":"https://doi.org/10.48550/arxiv.2605.13319","title":"PipeSD: An Efficient Cloud-Edge Collaborative Pipeline Inference Framework with Speculative Decoding","display_name":"PipeSD: An Efficient Cloud-Edge Collaborative Pipeline Inference Framework with Speculative Decoding","publication_year":2026,"publication_date":"2026-05-13","ids":{"openalex":"https://openalex.org/W7161115507","doi":"https://doi.org/10.48550/arxiv.2605.13319"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.13319","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.13319","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.13319","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136187419","display_name":"Yunhe Han","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han, Yunhe","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5011127712","display_name":"Yunqi Gao","orcid":"https://orcid.org/0009-0003-0599-8383"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Yunqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136169147","display_name":"Bing Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Bing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028862760","display_name":"Mahdi Boloursaz Mashhadi","orcid":"https://orcid.org/0000-0001-9948-9165"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mashhadi, Mahdi Boloursaz","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136101370","display_name":"Yitong Duan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Duan, Yitong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136100859","display_name":"Pei Xiao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiao, Pei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136120096","display_name":"Yanfeng Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yanfeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.3021000027656555,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.3021000027656555,"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":0.17229999601840973,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.05779999867081642,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.6740999817848206},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.6035000085830688},{"id":"https://openalex.org/keywords/cloud-computing","display_name":"Cloud computing","score":0.558899998664856},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5515999794006348},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.5236999988555908},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.5069000124931335},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.5041999816894531},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.4645000100135803}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8547999858856201},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6740999817848206},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.6035000085830688},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5860000252723694},{"id":"https://openalex.org/C79974875","wikidata":"https://www.wikidata.org/wiki/Q483639","display_name":"Cloud computing","level":2,"score":0.558899998664856},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5515999794006348},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.5236999988555908},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.5069000124931335},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.5041999816894531},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.4645000100135803},{"id":"https://openalex.org/C31395832","wikidata":"https://www.wikidata.org/wiki/Q1318674","display_name":"Testbed","level":2,"score":0.4611999988555908},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.36469998955726624},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.35740000009536743},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.33640000224113464},{"id":"https://openalex.org/C2780165032","wikidata":"https://www.wikidata.org/wiki/Q16869822","display_name":"Energy consumption","level":2,"score":0.3312999904155731},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.30070000886917114},{"id":"https://openalex.org/C2778476105","wikidata":"https://www.wikidata.org/wiki/Q628539","display_name":"Workload","level":2,"score":0.29809999465942383},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.29339998960494995},{"id":"https://openalex.org/C175309249","wikidata":"https://www.wikidata.org/wiki/Q725864","display_name":"Pipeline transport","level":2,"score":0.2858999967575073},{"id":"https://openalex.org/C2778156585","wikidata":"https://www.wikidata.org/wiki/Q174053","display_name":"Relay","level":3,"score":0.2529999911785126}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.13319","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.13319","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.13319","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.13319","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"display_name":"Affordable and clean energy","score":0.8891326189041138,"id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Speculative":[0],"decoding":[1,30],"can":[2],"significantly":[3],"accelerate":[4],"LLM":[5],"inference,":[6],"especially":[7],"given":[8],"that":[9,52,133],"its":[10],"cloud-edge":[11,67,121],"collaborative":[12,25,68],"deployment":[13],"offers":[14],"cloud":[15,47],"workload":[16],"offloading,":[17],"offline":[18],"robustness,":[19],"and":[20,38,44,79,91,113,115,142],"privacy":[21],"enhancement.":[22],"However,":[23],"existing":[24],"inference":[26,70],"frameworks":[27],"with":[28,40,72,101,123],"speculative":[29,73],"are":[31],"constrained":[32],"by":[33,81,88,146],"(i)":[34],"sequential":[35],"token":[36,77],"generation":[37,78],"communication":[39,80],"low":[41],"resource":[42],"utilization,":[43],"(ii)":[45],"inflexible":[46],"non-autoregressive":[48],"verification":[49,55,93],"(NAV)":[50],"triggering":[51,99],"induces":[53],"premature":[54],"or":[56],"costly":[57],"rollbacks.":[58],"In":[59],"this":[60],"paper,":[61],"we":[62],"propose":[63],"PipeSD,":[64],"an":[65],"efficient":[66],"pipeline":[69,84],"framework":[71],"decoding.":[74],"PipeSD":[75,109,134],"overlaps":[76],"a":[82,96,102,119],"token-batch":[83],"scheduling":[85],"mechanism":[86,100],"optimized":[87],"dynamic":[89],"programming,":[90],"improves":[92],"flexibility":[94],"through":[95],"dual-threshold":[97],"NAV":[98],"lightweight":[103],"Bayesian":[104],"optimization":[105],"autotuner.":[106],"We":[107],"implement":[108],"using":[110],"llama-cpp-python,":[111],"PyTorch,":[112],"FastAPI,":[114],"evaluate":[116],"it":[117],"on":[118],"real-world":[120],"testbed":[122],"two":[124],"draft-target":[125],"model":[126],"pairs":[127],"across":[128],"four":[129],"scenarios.":[130],"Results":[131],"show":[132],"consistently":[135],"outperforms":[136],"state-of-the-art":[137],"baselines,":[138],"achieving":[139],"1.16x-2.16x":[140],"speedup":[141],"reducing":[143],"energy":[144],"consumption":[145],"14.3%-25.3%.":[147]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-15T00:00:00"}