{"id":"https://openalex.org/W4416429987","doi":"https://doi.org/10.1109/iccad66269.2025.11240907","title":"SAGE: Saliency-Aware Grouping for Efficient Mapping of LLMs on Analog Compute-in-Memory","display_name":"SAGE: Saliency-Aware Grouping for Efficient Mapping of LLMs on Analog Compute-in-Memory","publication_year":2025,"publication_date":"2025-10-26","ids":{"openalex":"https://openalex.org/W4416429987","doi":"https://doi.org/10.1109/iccad66269.2025.11240907"},"language":null,"primary_location":{"id":"doi:10.1109/iccad66269.2025.11240907","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccad66269.2025.11240907","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/ACM International Conference On Computer Aided Design (ICCAD)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5014854872","display_name":"Y. Thomas Hou","orcid":"https://orcid.org/0009-0009-6780-076X"},"institutions":[{"id":"https://openalex.org/I165799507","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94","country_code":"US","type":"education","lineage":["https://openalex.org/I165799507"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yayue Hou","raw_affiliation_strings":["Rensselaer Polytechnic Institute,Troy,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Rensselaer Polytechnic Institute,Troy,USA","institution_ids":["https://openalex.org/I165799507"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100769921","display_name":"Zhenyu Liu","orcid":"https://orcid.org/0000-0002-9705-8377"},"institutions":[{"id":"https://openalex.org/I165799507","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94","country_code":"US","type":"education","lineage":["https://openalex.org/I165799507"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhenyu Liu","raw_affiliation_strings":["Rensselaer Polytechnic Institute,Troy,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Rensselaer Polytechnic Institute,Troy,USA","institution_ids":["https://openalex.org/I165799507"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119381438","display_name":"Garrett Gagnon","orcid":"https://orcid.org/0009-0009-0754-9518"},"institutions":[{"id":"https://openalex.org/I165799507","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94","country_code":"US","type":"education","lineage":["https://openalex.org/I165799507"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Garrett Gagnon","raw_affiliation_strings":["Rensselaer Polytechnic Institute,Troy,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Rensselaer Polytechnic Institute,Troy,USA","institution_ids":["https://openalex.org/I165799507"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086723230","display_name":"Hsinyu Tsai","orcid":"https://orcid.org/0000-0002-3971-097X"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hsinyu Tsai","raw_affiliation_strings":["IBM Research,San Jose,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Research,San Jose,USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034513925","display_name":"Kaoutar El Maghraoui","orcid":"https://orcid.org/0000-0002-1967-8749"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kaoutar El Maghraoui","raw_affiliation_strings":["IBM Research,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Research,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014981734","display_name":"Geoffrey W. Burr","orcid":"https://orcid.org/0000-0001-5717-2549"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Geoffrey W. Burr","raw_affiliation_strings":["IBM Research,San Jose,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IBM Research,San Jose,USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100324254","display_name":"Liu Liu","orcid":"https://orcid.org/0000-0001-8730-5824"},"institutions":[{"id":"https://openalex.org/I165799507","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94","country_code":"US","type":"education","lineage":["https://openalex.org/I165799507"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Liu Liu","raw_affiliation_strings":["Rensselaer Polytechnic Institute,Troy,USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Rensselaer Polytechnic Institute,Troy,USA","institution_ids":["https://openalex.org/I165799507"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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":"1","last_page":"9"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.3359000086784363,"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":0.3359000086784363,"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/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.3005000054836273,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.10679999738931656,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.5857999920845032},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.47909998893737793},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.44350001215934753},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4287000000476837},{"id":"https://openalex.org/keywords/statistical-inference","display_name":"Statistical inference","score":0.39800000190734863},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.39399999380111694},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.36000001430511475}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6654999852180481},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.5857999920845032},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.47909998893737793},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.44350001215934753},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4287000000476837},{"id":"https://openalex.org/C134261354","wikidata":"https://www.wikidata.org/wiki/Q938438","display_name":"Statistical inference","level":2,"score":0.39800000190734863},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.39399999380111694},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3898000121116638},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.36000001430511475},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.3422999978065491},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.30559998750686646},{"id":"https://openalex.org/C2742236","wikidata":"https://www.wikidata.org/wiki/Q924713","display_name":"Efficient energy use","level":2,"score":0.28439998626708984},{"id":"https://openalex.org/C164620267","wikidata":"https://www.wikidata.org/wiki/Q376953","display_name":"Adder","level":3,"score":0.2775000035762787},{"id":"https://openalex.org/C13412647","wikidata":"https://www.wikidata.org/wiki/Q174948","display_name":"Analog signal","level":3,"score":0.263700008392334},{"id":"https://openalex.org/C9083635","wikidata":"https://www.wikidata.org/wiki/Q2133535","display_name":"Noise shaping","level":2,"score":0.26109999418258667},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.260699987411499},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.260699987411499},{"id":"https://openalex.org/C205711294","wikidata":"https://www.wikidata.org/wiki/Q176953","display_name":"Rendering (computer graphics)","level":2,"score":0.260699987411499},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.25450000166893005}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/iccad66269.2025.11240907","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccad66269.2025.11240907","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/ACM International Conference On Computer Aided Design (ICCAD)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":36,"referenced_works":["https://openalex.org/W2508602506","https://openalex.org/W2518281301","https://openalex.org/W2613989746","https://openalex.org/W2963387357","https://openalex.org/W2964223283","https://openalex.org/W3003151806","https://openalex.org/W3005619596","https://openalex.org/W3042493405","https://openalex.org/W3176606713","https://openalex.org/W3191560475","https://openalex.org/W3213241618","https://openalex.org/W4210530509","https://openalex.org/W4214686755","https://openalex.org/W4220837341","https://openalex.org/W4285194374","https://openalex.org/W4286571874","https://openalex.org/W4292121737","https://openalex.org/W4313251083","https://openalex.org/W4360605785","https://openalex.org/W4366389026","https://openalex.org/W4376606760","https://openalex.org/W4385656513","https://openalex.org/W4385896568","https://openalex.org/W4386293192","https://openalex.org/W4390017976","https://openalex.org/W4390241252","https://openalex.org/W4391892592","https://openalex.org/W4392367648","https://openalex.org/W4393407269","https://openalex.org/W4399768602","https://openalex.org/W4402811198","https://openalex.org/W4404955344","https://openalex.org/W4406178148","https://openalex.org/W4406387997","https://openalex.org/W4406521098","https://openalex.org/W4410584352"],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"demand":[4],"high":[5],"memory":[6],"bandwidth":[7],"and":[8,31,49,57,77,101,123,149,159,168],"computational":[9],"efficiency,":[10],"posing":[11],"significant":[12],"challenges":[13],"for":[14,88],"deployment":[15],"on":[16,40,108,156,172],"traditional":[17],"digital":[18],"accelerators.":[19],"Analog":[20],"Compute-in-Memory":[21],"(ACIM)":[22],"architectures":[23],"offer":[24],"an":[25],"attractive":[26],"alternative":[27],"by":[28,98],"co-locating":[29],"storage":[30],"computation":[32],"to":[33,46,133],"reduce":[34],"data":[35,65],"movement.":[36],"However,":[37],"executing":[38],"LLMs":[39,158],"ACIM":[41,173],"systems":[42],"remains":[43],"challenging":[44],"due":[45],"hardware":[47],"non-idealities":[48],"the":[50,135],"unique":[51],"statistical":[52,109],"properties":[53],"of":[54,104,111],"LLM":[55],"inputs":[56],"outputs":[58],"in":[59],"FC":[60,105],"layers.":[61,154],"In":[62,80],"particular,":[63],"long-tailed":[64],"distributions":[66],"containing":[67],"large-amplitude":[68],"\"salient":[69],"values\"":[70],"degrade":[71],"analog":[72,121],"signal":[73],"quality":[74],"under":[75],"quantization":[76],"system":[78],"noise.":[79],"this":[81],"work,":[82],"we":[83],"propose":[84],"SAGE":[85,163],"(Saliency-Aware":[86],"Grouping":[87],"Efficient":[89],"Mapping),":[90],"a":[91,117,125,141],"training-free":[92],"strategy":[93],"that":[94,129,145,162],"improves":[95,165],"noise":[96],"resilience":[97],"reordering":[99],"weight":[100],"input":[102],"channels":[103],"layers":[106],"based":[107,171],"characteristics":[110],"LLMs.":[112],"We":[113,138],"identify":[114],"kurtosis":[115,132],"as":[116],"key":[118],"factor":[119],"affecting":[120],"robustness":[122],"develop":[124],"saliency-aware":[126],"mapping":[127],"method":[128],"reduces":[130],"output":[131],"enhance":[134],"signal-to-noise":[136],"ratio.":[137],"further":[139],"introduce":[140],"reconfigurable":[142],"tile":[143],"design":[144],"supports":[146],"mixed-precision":[147],"execution":[148],"maximizes":[150],"array":[151],"utilization":[152],"across":[153],"Evaluations":[155],"multiple":[157],"benchmarks":[160],"show":[161],"significantly":[164],"inference":[166],"accuracy":[167],"energy":[169],"efficiency":[170],"simulation":[174],"without":[175],"requiring":[176],"retraining.":[177]},"counts_by_year":[],"updated_date":"2026-07-15T18:14:33.161393","created_date":"2025-11-20T00:00:00"}