{"id":"https://openalex.org/W7166857337","doi":"https://doi.org/10.18653/v1/2026.findings-acl.230","title":"FlashMem: Distilling Intrinsic Latent Memory via Computation Reuse","display_name":"FlashMem: Distilling Intrinsic Latent Memory via Computation Reuse","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166857337","doi":"https://doi.org/10.18653/v1/2026.findings-acl.230"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.230","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.230","pdf_url":"https://aclanthology.org/2026.findings-acl.230.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.230.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139842282","display_name":"Yubo Hou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yubo Hou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139849531","display_name":"Zhisheng Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhisheng Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139795381","display_name":"Tao Wan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tao Wan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5032405950","display_name":"Zengchang Qin","orcid":"https://orcid.org/0000-0002-8084-6721"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zengchang Qin","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.84620704,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"4687","last_page":"4705"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.13680000603199005,"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":0.13680000603199005,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.09030000120401382,"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":0.06780000030994415,"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/computation","display_name":"Computation","score":0.4968999922275543},{"id":"https://openalex.org/keywords/reuse","display_name":"Reuse","score":0.44769999384880066},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.298799991607666},{"id":"https://openalex.org/keywords/component","display_name":"Component (thermodynamics)","score":0.29190000891685486}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6919999718666077},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4968999922275543},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.44769999384880066},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.33149999380111694},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.33009999990463257},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.298799991607666},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.29190000891685486},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.29030001163482666},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.26420000195503235},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.2393999993801117}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.230","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.230","pdf_url":"https://aclanthology.org/2026.findings-acl.230.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.230","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.230","pdf_url":"https://aclanthology.org/2026.findings-acl.230.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166857337.pdf","grobid_xml":"https://content.openalex.org/works/W7166857337.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"stateless":[1],"architecture":[2],"of":[3,90,136],"Large":[4],"Language":[5],"Models":[6],"inherently":[7],"lacks":[8],"the":[9,45,74,79,91,105,134,148],"mechanism":[10],"to":[11,17,21,98,104,119],"preserve":[12,70],"dynamic":[13],"context,":[14],"compelling":[15],"agents":[16],"redundantly":[18],"reprocess":[19],"history":[20],"maintain":[22],"longhorizon":[23],"autonomy.While":[24],"latent":[25],"memory":[26,43,55,100],"offers":[27],"a":[28,50,84,95,112],"solution,":[29],"current":[30],"approaches":[31],"are":[32],"hindered":[33],"by":[34,64,101,143],"architectural":[35],"segregation,":[36],"relying":[37],"on":[38],"auxiliary":[39],"encoders":[40],"that":[41,52,67,131],"decouple":[42],"from":[44,57],"reasoning":[46,59],"backbone.We":[47],"propose":[48],"FlashMem,":[49],"framework":[51],"distills":[53],"intrinsic":[54],"directly":[56,103],"transient":[58],"states":[60],"via":[61],"computation":[62],"reuse.Motivated":[63],"prior":[65],"findings":[66],"final-layer":[68],"representations":[69],"rich":[71],"information":[72],"about":[73],"input":[75],"trajectory,":[76],"Flash-Mem":[77],"uses":[78],"last":[80],"hidden":[81],"state":[82],"as":[83],"compact":[85],"and":[86,152],"empirically":[87],"effective":[88],"summary":[89],"interaction":[92],"history.This":[93],"enables":[94],"Shared-KV":[96],"Consolidator":[97],"synthesize":[99],"attending":[102],"backbone's":[106],"frozen":[107],"cache,":[108],"eliminating":[109],"redundant":[110],"re-parameterization.Furthermore,":[111],"parameter-free":[113],"Cognitive":[114],"Monitor":[115],"leverages":[116],"attention":[117],"entropy":[118],"adaptively":[120],"trigger":[121],"consolidation":[122],"only":[123],"when":[124],"high":[125],"epistemic":[126],"uncertainty":[127],"is":[128],"detected.Experiments":[129],"demonstrate":[130],"FlashMem":[132],"matches":[133],"performance":[135],"heavy":[137],"baselines":[138],"while":[139],"reducing":[140],"inference":[141],"latency":[142],"5":[144],"times,":[145],"effectively":[146],"bridging":[147],"gap":[149],"between":[150],"efficiency":[151],"persistent":[153],"cognition.":[154]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}