{"id":"https://openalex.org/W7125922553","doi":"https://doi.org/10.1109/smc58881.2025.11343559","title":"Dynamically Reconfigurable NPU Acceleration for Knowledge Loading in LLM Retrieval-Augmented Generation","display_name":"Dynamically Reconfigurable NPU Acceleration for Knowledge Loading in LLM Retrieval-Augmented Generation","publication_year":2025,"publication_date":"2025-10-05","ids":{"openalex":"https://openalex.org/W7125922553","doi":"https://doi.org/10.1109/smc58881.2025.11343559"},"language":null,"primary_location":{"id":"doi:10.1109/smc58881.2025.11343559","is_oa":false,"landing_page_url":"https://doi.org/10.1109/smc58881.2025.11343559","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Conference on Systems, Man, and Cybernetics (SMC)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5124068500","display_name":"Peidong Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Peidong Lin","raw_affiliation_strings":["Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110"],"affiliations":[{"raw_affiliation_string":"Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124120582","display_name":"Jintao Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jintao Li","raw_affiliation_strings":["Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110"],"affiliations":[{"raw_affiliation_string":"Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124056433","display_name":"Hui Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hui Deng","raw_affiliation_strings":["Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110"],"affiliations":[{"raw_affiliation_string":"Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124068313","display_name":"Shihong Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shihong Li","raw_affiliation_strings":["Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110"],"affiliations":[{"raw_affiliation_string":"Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5008923294","display_name":"Shui Yu","orcid":"https://orcid.org/0000-0003-3473-5032"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shui Yu","raw_affiliation_strings":["Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110"],"affiliations":[{"raw_affiliation_string":"Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5124083011","display_name":"Yun Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yun Li","raw_affiliation_strings":["Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110"],"affiliations":[{"raw_affiliation_string":"Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China,Shenzhen,China,518110","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5124068500"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.83878316,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"2685","last_page":"2690"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.5439000129699707,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.5439000129699707,"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/T10028","display_name":"Topic Modeling","score":0.12870000302791595,"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/T14347","display_name":"Big Data and Digital Economy","score":0.09220000356435776,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/speedup","display_name":"Speedup","score":0.7249000072479248},{"id":"https://openalex.org/keywords/parsing","display_name":"Parsing","score":0.6287999749183655},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6147000193595886},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5314000248908997},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.5058000087738037},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.4546999931335449},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.42289999127388},{"id":"https://openalex.org/keywords/energy","display_name":"Energy (signal processing)","score":0.42100000381469727}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8116000294685364},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.7249000072479248},{"id":"https://openalex.org/C186644900","wikidata":"https://www.wikidata.org/wiki/Q194152","display_name":"Parsing","level":2,"score":0.6287999749183655},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6147000193595886},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5314000248908997},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5058000087738037},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.4546999931335449},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.42289999127388},{"id":"https://openalex.org/C186370098","wikidata":"https://www.wikidata.org/wiki/Q442787","display_name":"Energy (signal processing)","level":2,"score":0.42100000381469727},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.4018000066280365},{"id":"https://openalex.org/C60690694","wikidata":"https://www.wikidata.org/wiki/Q894902","display_name":"Bottom-up parsing","level":4,"score":0.3837999999523163},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.36070001125335693},{"id":"https://openalex.org/C115901376","wikidata":"https://www.wikidata.org/wiki/Q184199","display_name":"Automation","level":2,"score":0.3596000075340271},{"id":"https://openalex.org/C2780165032","wikidata":"https://www.wikidata.org/wiki/Q16869822","display_name":"Energy consumption","level":2,"score":0.35690000653266907},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.35010001063346863},{"id":"https://openalex.org/C9390403","wikidata":"https://www.wikidata.org/wiki/Q3966","display_name":"Computer hardware","level":1,"score":0.3368000090122223},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3131999969482422},{"id":"https://openalex.org/C206345919","wikidata":"https://www.wikidata.org/wiki/Q20380951","display_name":"Resource (disambiguation)","level":2,"score":0.3091000020503998},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2930999994277954},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.2806999981403351},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.26969999074935913},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.26649999618530273},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.258899986743927},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.25429999828338623},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.25380000472068787},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.25290000438690186}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/smc58881.2025.11343559","is_oa":false,"landing_page_url":"https://doi.org/10.1109/smc58881.2025.11343559","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Conference on Systems, Man, and Cybernetics (SMC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.9015081524848938,"display_name":"Affordable and clean energy","id":"https://metadata.un.org/sdg/7"}],"awards":[],"funders":[{"id":"https://openalex.org/F4320337504","display_name":"Research and Development","ror":"https://ror.org/027s68j25"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":29,"referenced_works":["https://openalex.org/W2080633243","https://openalex.org/W2083116240","https://openalex.org/W2092619708","https://openalex.org/W2104322476","https://openalex.org/W2124509324","https://openalex.org/W2940553501","https://openalex.org/W2990397035","https://openalex.org/W3004829073","https://openalex.org/W3016939927","https://openalex.org/W3190761184","https://openalex.org/W4214950035","https://openalex.org/W4255133361","https://openalex.org/W4281673407","https://openalex.org/W4309674289","https://openalex.org/W4317898419","https://openalex.org/W4384947530","https://openalex.org/W4386474956","https://openalex.org/W4387321091","https://openalex.org/W4389162407","https://openalex.org/W4389776171","https://openalex.org/W4390337049","https://openalex.org/W4391667303","https://openalex.org/W4393146755","https://openalex.org/W4401042753","https://openalex.org/W4402742645","https://openalex.org/W4407197060","https://openalex.org/W4408019906","https://openalex.org/W4408151489","https://openalex.org/W4409282428"],"related_works":[],"abstract_inverted_index":{"Retrieval-Augmented":[0],"Generation":[1],"(RAG)":[2],"provides":[3],"large":[4],"language":[5],"models":[6],"(LLMs)":[7],"a":[8,32,74,81,97,108],"means":[9],"of":[10],"retrieving":[11],"relevant":[12],"external":[13],"knowledge,":[14],"but":[15],"its":[16],"document":[17,43,78],"parsing":[18,44,79],"leads":[19],"to":[20],"increased":[21],"latency":[22],"and":[23,45,53,62,80,96],"energy":[24,85,101],"consumption.":[25],"To":[26],"address":[27],"this":[28],"issue,":[29],"we":[30],"propose":[31],"dynamically":[33],"reconfigurable":[34],"Neural":[35],"Processing":[36],"Unit":[37],"(NPU)":[38],"that":[39,70],"accelerates":[40],"both":[41],"RAG":[42,113],"inference.":[46],"By":[47],"leveraging":[48],"compute-in-memory":[49],"fusion,":[50],"dynamic":[51],"convolution,":[52],"multi-level":[54],"parallelism,":[55],"our":[56,71],"approach":[57],"reduces":[58],"memory":[59],"transfer":[60],"overhead":[61],"optimizes":[63],"hardware":[64],"resource":[65],"allocation.":[66],"Experimental":[67],"results":[68],"show":[69],"design":[72],"achieves":[73,89],"1.8x":[75],"speedup":[76],"in":[77,84,93,100],"2.83x":[82],"improvement":[83],"efficiency.":[86],"Additionally,":[87],"it":[88],"an":[90],"11.71%":[91],"reduction":[92],"inference":[94],"time":[95],"35.59%":[98],"boost":[99],"efficiency":[102],"over":[103],"traditional":[104],"CPU/GPU":[105],"methods,":[106],"offering":[107],"scalable":[109],"solution":[110],"for":[111],"large-scale":[112],"tasks.":[114]},"counts_by_year":[],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2026-01-29T00:00:00"}