{"id":"https://openalex.org/W7168367751","doi":"https://doi.org/10.48448/c2hj-3z83","title":"Lightweight LLM Agent Memory with Small Language Models","display_name":"Lightweight LLM Agent Memory with Small Language Models","publication_year":2026,"publication_date":"2026-06-03","ids":{"openalex":"https://openalex.org/W7168367751","doi":"https://doi.org/10.48448/c2hj-3z83"},"language":"en","primary_location":{"id":"doi:10.48448/c2hj-3z83","doi":"https://doi.org/10.48448/c2hj-3z83","is_oa":true,"landing_page_url":"https://doi.org/10.48448/c2hj-3z83","pdf_url":null,"source":{"id":"https://openalex.org/S7407053148","display_name":"Underline Science Inc.","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Audiovisual"},"volume":null,"type":"other","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48448/c2hj-3z83","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5140798803","display_name":"Association for Computational Linguistics 2026","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Association for Computational Linguistics 2026","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125760211","display_name":"Sung-Ho Bae","orcid":"https://orcid.org/0000-0002-3389-1159","observed_orcids":["https://orcid.org/0000-0002-3389-1159","https://orcid.org/0000-0003-2677-3186"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bae, Sung-Ho","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140909140","display_name":"Shuxu Chen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Shuxu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069989561","display_name":"Ping Guo","orcid":"https://orcid.org/0000-0001-5412-914X","observed_orcids":["https://orcid.org/0000-0001-5412-914X","https://orcid.org/0009-0005-1166-4042"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Ping","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134781215","display_name":"Zhenzhen Huang","orcid":"https://orcid.org/0009-0000-3518-6130","observed_orcids":["https://orcid.org/0009-0000-3518-6130"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Zhenzhen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140784814","display_name":"Fan Mo","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mo, Fan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121484045","display_name":"Zhicheng Wang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Zhicheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140802513","display_name":"Jiwei Wei","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei, Jiwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140910235","display_name":"Yang Yang","orcid":"https://orcid.org/0000-0002-5070-4511","observed_orcids":["https://orcid.org/0000-0002-5070-4511"]},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Yang","raw_affiliation_strings":["University of Electronic Science and Technology of China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057230698","display_name":"Chaoning Zhang","orcid":"https://orcid.org/0000-0001-6007-6099","observed_orcids":["https://orcid.org/0000-0001-6007-6099","https://orcid.org/0000-0003-3484-7333"]},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Chaoning","raw_affiliation_strings":["University of Electronic Science and Technology of China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140880227","display_name":"Jiaquan Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Jiaquan","raw_affiliation_strings":["University of Electronic Science and Technology of China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123929105","display_name":"Pengcheng Zheng","orcid":"https://orcid.org/0009-0004-3323-9170","observed_orcids":["https://orcid.org/0009-0004-3323-9170"]},"institutions":[{"id":"https://openalex.org/I150229711","display_name":"University of Electronic Science and Technology of China","ror":"https://ror.org/04qr3zq92","country_code":"CN","type":"education","lineage":["https://openalex.org/I150229711"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zheng, Pengcheng","raw_affiliation_strings":["University of Electronic Science and Technology of China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Electronic Science and Technology of China","institution_ids":["https://openalex.org/I150229711"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5044839127","display_name":"Jie Zou","orcid":"https://orcid.org/0000-0002-0181-778X","observed_orcids":["https://orcid.org/0000-0002-0181-778X","https://orcid.org/0000-0003-0911-8542"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zou, Jie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"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/T10028","display_name":"Topic Modeling","score":0.38089999556541443,"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/T10028","display_name":"Topic Modeling","score":0.38089999556541443,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.298799991607666,"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"}},{"id":"https://openalex.org/T12031","display_name":"Speech and dialogue systems","score":0.10170000046491623,"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/llm-based-agents","display_name":"LLM agents","score":0.9049999713897705},{"id":"https://openalex.org/keywords/agent-memory","display_name":"agent memory","score":0.9399999976158142},{"id":"https://openalex.org/keywords/small-language-models","display_name":"small language models","score":0.9679999947547913},{"id":"https://openalex.org/keywords/long-term-memory-consolidation","display_name":"long-term memory consolidation","score":0.7310000061988831},{"id":"https://openalex.org/keywords/short-term-memory","display_name":"short-term memory","score":0.7730000019073486},{"id":"https://openalex.org/keywords/retrieval-augmented-generation","display_name":"retrieval-augmented generation","score":0.6420000195503235},{"id":"https://openalex.org/keywords/long-horizon-interaction","display_name":"long-horizon interaction","score":0.5960000157356262}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8331999778747559},{"id":"https://openalex.org/C12186640","wikidata":"https://www.wikidata.org/wiki/Q6815743","display_name":"Memory model","level":3,"score":0.5430999994277954},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5163999795913696},{"id":"https://openalex.org/C37279795","wikidata":"https://www.wikidata.org/wiki/Q2492305","display_name":"Consistency model","level":3,"score":0.5037999749183655},{"id":"https://openalex.org/C197914299","wikidata":"https://www.wikidata.org/wiki/Q18650","display_name":"Semantic memory","level":3,"score":0.503000020980835},{"id":"https://openalex.org/C74426580","wikidata":"https://www.wikidata.org/wiki/Q719484","display_name":"Memory map","level":3,"score":0.45989999175071716},{"id":"https://openalex.org/C136085584","wikidata":"https://www.wikidata.org/wiki/Q910289","display_name":"Overlay","level":2,"score":0.42750000953674316},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.41850000619888306},{"id":"https://openalex.org/C15569618","wikidata":"https://www.wikidata.org/wiki/Q3561421","display_name":"Liveness","level":2,"score":0.40149998664855957},{"id":"https://openalex.org/C48455012","wikidata":"https://www.wikidata.org/wiki/Q2892593","display_name":"Memory consolidation","level":3,"score":0.39169999957084656},{"id":"https://openalex.org/C39528615","wikidata":"https://www.wikidata.org/wiki/Q1229610","display_name":"Distributed shared memory","level":5,"score":0.38029998540878296},{"id":"https://openalex.org/C133875982","wikidata":"https://www.wikidata.org/wiki/Q764810","display_name":"Shared memory","level":2,"score":0.37689998745918274},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.3765000104904175},{"id":"https://openalex.org/C57863822","wikidata":"https://www.wikidata.org/wiki/Q905488","display_name":"Flat memory model","level":4,"score":0.3707999885082245},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.3560999929904938},{"id":"https://openalex.org/C141917322","wikidata":"https://www.wikidata.org/wiki/Q1025017","display_name":"Cache coherence","level":5,"score":0.3458000123500824},{"id":"https://openalex.org/C63511323","wikidata":"https://www.wikidata.org/wiki/Q908936","display_name":"Interleaved memory","level":4,"score":0.3418000042438507},{"id":"https://openalex.org/C175652121","wikidata":"https://www.wikidata.org/wiki/Q4379351","display_name":"Causal consistency","level":5,"score":0.3377000093460083},{"id":"https://openalex.org/C154504017","wikidata":"https://www.wikidata.org/wiki/Q853614","display_name":"Identifier","level":2,"score":0.3361999988555908},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.3353999853134155},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.3296999931335449},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.30469998717308044},{"id":"https://openalex.org/C5165142","wikidata":"https://www.wikidata.org/wiki/Q5432732","display_name":"False sharing","level":5,"score":0.2962000072002411},{"id":"https://openalex.org/C112049663","wikidata":"https://www.wikidata.org/wiki/Q18608","display_name":"Explicit memory","level":4,"score":0.2922999858856201},{"id":"https://openalex.org/C51290061","wikidata":"https://www.wikidata.org/wiki/Q1936765","display_name":"Uniform memory access","level":4,"score":0.29179999232292175},{"id":"https://openalex.org/C171675096","wikidata":"https://www.wikidata.org/wiki/Q1143380","display_name":"Extended memory","level":4,"score":0.28870001435279846},{"id":"https://openalex.org/C119907115","wikidata":"https://www.wikidata.org/wiki/Q6815725","display_name":"Memory errors","level":3,"score":0.2791000008583069},{"id":"https://openalex.org/C82029504","wikidata":"https://www.wikidata.org/wiki/Q4373882","display_name":"Sequential consistency","level":4,"score":0.2782999873161316},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.27630001306533813},{"id":"https://openalex.org/C93446704","wikidata":"https://www.wikidata.org/wiki/Q449328","display_name":"Registered memory","level":3,"score":0.27570000290870667},{"id":"https://openalex.org/C120039088","wikidata":"https://www.wikidata.org/wiki/Q6875176","display_name":"Misattribution of memory","level":3,"score":0.275299996137619},{"id":"https://openalex.org/C53838383","wikidata":"https://www.wikidata.org/wiki/Q541148","display_name":"Conventional memory","level":5,"score":0.2750000059604645},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.2612000107765198},{"id":"https://openalex.org/C122377713","wikidata":"https://www.wikidata.org/wiki/Q4422799","display_name":"Weak consistency","level":4,"score":0.2590999901294708},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C88576662","wikidata":"https://www.wikidata.org/wiki/Q18646","display_name":"Episodic memory","level":3,"score":0.25540000200271606}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48448/c2hj-3z83","doi":"https://doi.org/10.48448/c2hj-3z83","is_oa":true,"landing_page_url":"https://doi.org/10.48448/c2hj-3z83","pdf_url":null,"source":{"id":"https://openalex.org/S7407053148","display_name":"Underline Science Inc.","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Audiovisual"}],"best_oa_location":{"id":"doi:10.48448/c2hj-3z83","doi":"https://doi.org/10.48448/c2hj-3z83","is_oa":true,"landing_page_url":"https://doi.org/10.48448/c2hj-3z83","pdf_url":null,"source":{"id":"https://openalex.org/S7407053148","display_name":"Underline Science Inc.","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Audiovisual"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4416057654","https://openalex.org/W7168404910","https://openalex.org/W4416361029","https://openalex.org/W7216504145","https://openalex.org/W7161581122","https://openalex.org/W7118586615","https://openalex.org/W7172528580","https://openalex.org/W7220767060","https://openalex.org/W7168616902","https://openalex.org/W7171733833"],"abstract_inverted_index":{"Although":[0],"LLM":[1],"agents":[2],"can":[3],"leverage":[4],"tools":[5],"for":[6,54,73,113,120,128,209],"complex":[7],"tasks,":[8],"they":[9],"still":[10],"need":[11],"memory":[12,27,56,71,76,85,101,108,111,118,126],"to":[13,39,98,135],"maintain":[14],"cross-turn":[15],"consistency":[16,166],"and":[17,43,88,91,124,131,139,153,174,203,211],"accumulate":[18],"reusable":[19,121,171],"information":[20],"in":[21,142],"long-horizon":[22],"interactions.":[23,65],"However,":[24],"retrieval-based":[25],"external":[26],"systems":[28,49],"incur":[29],"low":[30,204],"online":[31,55,93],"overhead":[32],"but":[33,60],"suffer":[34],"from":[35,95],"unstable":[36],"accuracy":[37,59],"due":[38],"limited":[40],"query":[41],"construction":[42],"candidate":[44],"filtering.":[45],"In":[46],"contrast,":[47],"many":[48],"use":[50],"repeated":[51],"large-model":[52],"calls":[53],"operations,":[57],"improving":[58],"accumulating":[61],"latency":[62,206],"over":[63,195],"long":[64],"We":[66,106],"propose":[67],"LightMem,":[68],"a":[69,149,157],"lightweight":[70],"system":[72],"better":[74],"agent":[75],"driven":[77],"by":[78,164],"Small":[79],"Language":[80],"Models":[81],"(SLMs).":[82],"LightMem":[83,146],"modularizes":[84],"retrieval,":[86],"writing,":[87],"long-term":[89,125],"consolidation,":[90],"separates":[92],"processing":[94],"offline":[96],"consolidation":[97],"enable":[99],"efficient":[100],"invocation":[102],"under":[103,148],"bounded":[104],"compute.":[105],"organize":[107],"into":[109,178],"short-term":[110],"(STM)":[112],"immediate":[114],"conversational":[115],"context,":[116],"mid-term":[117],"(MTM)":[119],"interaction":[122,172],"summaries,":[123],"(LTM)":[127],"consolidated":[129],"knowledge,":[130],"uses":[132],"user":[133],"identifiers":[134],"support":[136],"independent":[137],"retrieval":[138,151,162,210],"incremental":[140],"maintenance":[141],"multi-user":[143],"settings.":[144],"Online,":[145],"operates":[147],"fixed":[150],"budget":[152],"selects":[154],"memories":[155],"via":[156],"two-stage":[158],"procedure:":[159],"vector-based":[160],"coarse":[161],"followed":[163],"semantic":[165],"re-ranking.":[167],"Offline,":[168],"it":[169,177],"abstracts":[170],"evidence":[173],"incrementally":[175],"integrates":[176],"LTM.":[179],"Experiments":[180],"show":[181],"consistent":[182],"gains":[183],"across":[184],"model":[185],"scales,":[186],"with":[187],"an":[188],"average":[189],"F1":[190],"improvement":[191],"of":[192],"about":[193],"2.5":[194],"A-MEM":[196],"on":[197],"LoCoMo,":[198],"while":[199],"achieving":[200],"higher":[201],"efficiency":[202],"median":[205],"(83":[207],"ms":[208,213],"581":[212],"end-to-end).":[214]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-07-16T00:00:00"}