{"id":"https://openalex.org/W7151296371","doi":"https://doi.org/10.48550/arxiv.2604.03679","title":"LightThinker++: From Reasoning Compression to Memory Management","display_name":"LightThinker++: From Reasoning Compression to Memory Management","publication_year":2026,"publication_date":"2026-04-04","ids":{"openalex":"https://openalex.org/W7151296371","doi":"https://doi.org/10.48550/arxiv.2604.03679"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.03679","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.03679","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":null,"license_id":null,"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.2604.03679","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5133134130","display_name":"Yuqi Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Yuqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133134045","display_name":"Jintian Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Jintian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133140975","display_name":"Zhenjie Wan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wan, Zhenjie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133129709","display_name":"Yujie Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Yujie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133086301","display_name":"Shuofei Qiao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qiao, Shuofei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133091185","display_name":"Zhengke Gui","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gui, Zhengke","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133096970","display_name":"Da Wei Zheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Da","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133122559","display_name":"Lei Liang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liang, Lei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133107154","display_name":"Huajun Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Huajun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133143939","display_name":"Ningyu Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Ningyu","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":0.2531999945640564,"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":0.2531999945640564,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.20229999721050262,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T13702","display_name":"Machine Learning in Healthcare","score":0.08950000256299973,"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/context","display_name":"Context (archaeology)","score":0.5687999725341797},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.5396999716758728},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5302000045776367},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.5284000039100647},{"id":"https://openalex.org/keywords/memory-footprint","display_name":"Memory footprint","score":0.5174000263214111},{"id":"https://openalex.org/keywords/memory-management","display_name":"Memory management","score":0.491100013256073},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.47999998927116394},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.4253000020980835},{"id":"https://openalex.org/keywords/memory-model","display_name":"Memory model","score":0.4000999927520752}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.796500027179718},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5687999725341797},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5396999716758728},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5302000045776367},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.5284000039100647},{"id":"https://openalex.org/C74912251","wikidata":"https://www.wikidata.org/wiki/Q6815727","display_name":"Memory footprint","level":2,"score":0.5174000263214111},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.491100013256073},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.47999998927116394},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.4253000020980835},{"id":"https://openalex.org/C12186640","wikidata":"https://www.wikidata.org/wiki/Q6815743","display_name":"Memory model","level":3,"score":0.4000999927520752},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.3912999927997589},{"id":"https://openalex.org/C169900460","wikidata":"https://www.wikidata.org/wiki/Q2200417","display_name":"Cognition","level":2,"score":0.37880000472068787},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.37689998745918274},{"id":"https://openalex.org/C81081738","wikidata":"https://www.wikidata.org/wiki/Q55542","display_name":"Lossless compression","level":3,"score":0.3671000003814697},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3601999878883362},{"id":"https://openalex.org/C100660578","wikidata":"https://www.wikidata.org/wiki/Q18733","display_name":"Recall","level":2,"score":0.3434999883174896},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.34139999747276306},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.32440000772476196},{"id":"https://openalex.org/C160145156","wikidata":"https://www.wikidata.org/wiki/Q778586","display_name":"Executable","level":2,"score":0.3093000054359436},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.30630001425743103},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.287200003862381},{"id":"https://openalex.org/C197914299","wikidata":"https://www.wikidata.org/wiki/Q18650","display_name":"Semantic memory","level":3,"score":0.2782000005245209},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.274399995803833},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.27309998869895935},{"id":"https://openalex.org/C132943942","wikidata":"https://www.wikidata.org/wiki/Q2562511","display_name":"Footprint","level":2,"score":0.2718999981880188},{"id":"https://openalex.org/C171675096","wikidata":"https://www.wikidata.org/wiki/Q1143380","display_name":"Extended memory","level":4,"score":0.27070000767707825},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.2696000039577484},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.2660999894142151},{"id":"https://openalex.org/C129916263","wikidata":"https://www.wikidata.org/wiki/Q1141183","display_name":"Backward chaining","level":4,"score":0.2563000023365021}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.03679","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.03679","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.03679","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.03679","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2],"(LLMs)":[3],"excel":[4],"at":[5],"complex":[6,48,177],"reasoning,":[7,128],"yet":[8],"their":[9],"efficiency":[10],"is":[11],"limited":[12],"by":[13,82,88,114,119,134],"the":[14,51,67,102,143],"surging":[15],"cognitive":[16],"overhead":[17],"of":[18,54,173],"long":[19],"thought":[20],"traces.":[21],"In":[22,126],"this":[23],"paper,":[24],"we":[25,65],"propose":[26],"LightThinker,":[27],"a":[28,89,138,159,183],"method":[29],"that":[30],"enables":[31],"LLMs":[32],"to":[33,59,79,94],"dynamically":[34],"compress":[35],"intermediate":[36,55],"thoughts":[37],"into":[38,69],"compact":[39],"semantic":[40],"representations.":[41],"However,":[42],"static":[43],"compression":[44],"often":[45],"struggles":[46],"with":[47,121,194],"reasoning":[49,190],"where":[50],"irreversible":[52],"loss":[53],"details":[56],"can":[57],"lead":[58],"logical":[60],"bottlenecks.":[61],"To":[62],"address":[63],"this,":[64],"evolve":[66],"framework":[68],"LightThinker++,":[70],"introducing":[71],"Explicit":[72],"Adaptive":[73],"Memory":[74],"Management.":[75],"This":[76],"paradigm":[77],"shifts":[78],"behavioral-level":[80],"management":[81],"incorporating":[83],"explicit":[84],"memory":[85,97],"primitives,":[86],"supported":[87],"specialized":[90],"trajectory":[91],"synthesis":[92],"pipeline":[93],"train":[95],"purposeful":[96],"scheduling.":[98],"Extensive":[99],"experiments":[100],"demonstrate":[101],"framework's":[103],"versatility":[104],"across":[105,175],"three":[106],"dimensions.":[107],"(1)":[108],"LightThinker":[109],"reduces":[110],"peak":[111,131],"token":[112,132],"usage":[113,133],"70%":[115],"and":[116],"inference":[117],"time":[118],"26%":[120],"minimal":[122,195],"accuracy":[123,140],"loss.":[124],"(2)":[125],"standard":[127],"LightThinker++":[129],"slashes":[130],"69.9%":[135],"while":[136],"yielding":[137],"+2.42%":[139],"gain":[141,172],"under":[142],"same":[144],"context":[145],"budget":[146],"for":[147,186],"maximum":[148],"performance.":[149],"(3)":[150],"Most":[151],"notably,":[152],"in":[153],"long-horizon":[154],"agentic":[155],"tasks,":[156],"it":[157],"maintains":[158],"stable":[160],"footprint":[161],"beyond":[162],"80":[163],"rounds":[164],"(a":[165],"60%-70%":[166],"reduction),":[167],"achieving":[168],"an":[169],"average":[170],"performance":[171],"14.8%":[174],"different":[176],"scenarios.":[178],"Overall,":[179],"our":[180],"work":[181],"provides":[182],"scalable":[184],"direction":[185],"sustaining":[187],"deep":[188],"LLM":[189],"over":[191],"extended":[192],"horizons":[193],"overhead.":[196]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-08T00:00:00"}