{"id":"https://openalex.org/W7160116345","doi":"https://doi.org/10.48550/arxiv.2605.00356","title":"MemRouter: Memory-as-Embedding Routing for Long-Term Conversational Agents","display_name":"MemRouter: Memory-as-Embedding Routing for Long-Term Conversational Agents","publication_year":2026,"publication_date":"2026-05-01","ids":{"openalex":"https://openalex.org/W7160116345","doi":"https://doi.org/10.48550/arxiv.2605.00356"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.00356","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.00356","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2605.00356","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5059058846","display_name":"Tianyu Hu","orcid":"https://orcid.org/0000-0001-9903-0696"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Tianyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135150559","display_name":"Weikai Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Weikai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135128370","display_name":"Weizhi Zhang","orcid":"https://orcid.org/0009-0006-7613-6774"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Weizhi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135103585","display_name":"Jing Ma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Jing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135177967","display_name":"Song Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Song","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/T10028","display_name":"Topic Modeling","score":0.8382999897003174,"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":0.8382999897003174,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.03799999877810478,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.019899999722838402,"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/latency","display_name":"Latency (audio)","score":0.651199996471405},{"id":"https://openalex.org/keywords/router","display_name":"Router","score":0.5473999977111816},{"id":"https://openalex.org/keywords/downstream","display_name":"Downstream (manufacturing)","score":0.5343000292778015},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.524399995803833},{"id":"https://openalex.org/keywords/component","display_name":"Component (thermodynamics)","score":0.4848000109195709},{"id":"https://openalex.org/keywords/encoding","display_name":"Encoding (memory)","score":0.4189999997615814},{"id":"https://openalex.org/keywords/routing","display_name":"Routing (electronic design automation)","score":0.40059998631477356}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7336000204086304},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.651199996471405},{"id":"https://openalex.org/C2775896111","wikidata":"https://www.wikidata.org/wiki/Q642560","display_name":"Router","level":2,"score":0.5473999977111816},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.5343000292778015},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.524399995803833},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.4848000109195709},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.42899999022483826},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.4189999997615814},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.40059998631477356},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.36480000615119934},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.35929998755455017},{"id":"https://openalex.org/C554936623","wikidata":"https://www.wikidata.org/wiki/Q199657","display_name":"Reading (process)","level":2,"score":0.31869998574256897},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.311599999666214},{"id":"https://openalex.org/C2984173633","wikidata":"https://www.wikidata.org/wiki/Q22725","display_name":"Routing algorithm","level":4,"score":0.3095000088214874},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3066999912261963},{"id":"https://openalex.org/C66746571","wikidata":"https://www.wikidata.org/wiki/Q1134833","display_name":"ENCODE","level":3,"score":0.295199990272522},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.2892000079154968},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.26190000772476196},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.25519999861717224},{"id":"https://openalex.org/C169258074","wikidata":"https://www.wikidata.org/wiki/Q245748","display_name":"Random forest","level":2,"score":0.25279998779296875}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.00356","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.00356","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.00356","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.00356","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions","score":0.702858567237854}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Long-term":[0],"conversational":[1,187],"agents":[2],"must":[3],"decide":[4],"which":[5],"turns":[6],"to":[7,23,132],"store":[8],"in":[9,185],"external":[10],"memory,":[11],"yet":[12],"recent":[13,59],"systems":[14],"rely":[15],"on":[16,92,113],"autoregressive":[17],"LLM":[18,68],"generation":[19,179],"at":[20],"every":[21,114],"turn":[22,56,74],"make":[24],"that":[25,34,139,165],"decision.":[26],"We":[27],"present":[28],"MemRouter,":[29],"a":[30,66,88,155,173,181],"write-side":[31,166],"memory":[32,36,111,167],"router":[33],"decouples":[35],"admission":[37,141,168],"from":[38,130],"the":[39,62,73,95],"downstream":[40,183],"answer":[41,98,178],"backbone":[42,102],"and":[43,70,100,158],"replaces":[44],"per-turn":[45],"memory-management":[46,127],"decoding":[47],"with":[48,58],"an":[49,109],"embedding-based":[50],"routing":[51],"policy.":[52],"MemRouter":[53,107],"encodes":[54],"each":[55],"together":[57],"context,":[60],"projects":[61],"resulting":[63],"embeddings":[64],"through":[65],"frozen":[67],"backbone,":[69],"predicts":[71],"whether":[72],"should":[75],"be":[76,170],"stored":[77],"using":[78],"lightweight":[79],"classification":[80],"heads":[81],"while":[82,125,177],"training":[83],"only":[84],"12M":[85],"parameters.":[86],"Under":[87],"controlled":[89],"matched-harness":[90],"comparison":[91],"LoCoMo,":[93],"where":[94],"retrieval":[96,159],"pipeline,":[97],"prompts,":[99],"QA":[101],"(Qwen2.5-7B)":[103],"are":[104],"held":[105],"identical,":[106],"outperforms":[108],"LLM-based":[110],"manager":[112],"question":[115],"category":[116],"(overall":[117],"F1":[118,144],"52.0":[119],"vs":[120],"45.6,":[121],"non-overlapping":[122],"95%":[123],"CIs)":[124],"reducing":[126],"p50":[128],"latency":[129],"970ms":[131],"58ms.":[133],"Descriptive":[134],"factorial":[135],"averaging":[136],"further":[137],"shows":[138],"learned":[140,171],"improves":[142],"mean":[143],"by":[145,172],"+10.3":[146],"over":[147,154],"random":[148],"storage,":[149],"category-specific":[150],"prompting":[151],"adds":[152],"+5.2":[153],"generic":[156],"prompt,":[157],"contributes":[160],"+0.7.":[161],"These":[162],"results":[163],"suggest":[164],"can":[169],"small":[174],"supervised":[175],"router,":[176],"remains":[180],"separate":[182],"component":[184],"long-horizon":[186],"QA.":[188]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-05T00:00:00"}