{"id":"https://openalex.org/W7130348822","doi":"https://doi.org/10.48550/arxiv.2602.15156","title":"Panini: Continual Learning in Token Space via Structured Memory","display_name":"Panini: Continual Learning in Token Space via Structured Memory","publication_year":2026,"publication_date":"2026-02-16","ids":{"openalex":"https://openalex.org/W7130348822","doi":"https://doi.org/10.48550/arxiv.2602.15156"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2602.15156","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.15156","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.2602.15156","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5088303386","display_name":"Shreyas Rajesh","orcid":"https://orcid.org/0009-0002-8077-530X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rajesh, Shreyas","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057287572","display_name":"Pavan Holur","orcid":"https://orcid.org/0000-0002-8495-9416"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Holur, Pavan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092812154","display_name":"Mehmet Y. Turali","orcid":"https://orcid.org/0000-0002-6147-1741"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Turali, Mehmet Yigit","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054458304","display_name":"Chenda Duan","orcid":"https://orcid.org/0009-0003-8652-3960"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Duan, Chenda","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5126313425","display_name":"Vwani Roychowdhury","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Roychowdhury, Vwani","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":0.7103000283241272,"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":0.7103000283241272,"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":0.09860000014305115,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.01889999955892563,"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/inference","display_name":"Inference","score":0.7612000107765198},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.59170001745224},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.5180000066757202},{"id":"https://openalex.org/keywords/structuring","display_name":"Structuring","score":0.5044999718666077},{"id":"https://openalex.org/keywords/reliability","display_name":"Reliability (semiconductor)","score":0.4918000102043152},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.4681999981403351},{"id":"https://openalex.org/keywords/structured-prediction","display_name":"Structured prediction","score":0.4404999911785126},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.39899998903274536},{"id":"https://openalex.org/keywords/space","display_name":"Space (punctuation)","score":0.3977999985218048}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8179000020027161},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7612000107765198},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.59170001745224},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5180000066757202},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5120999813079834},{"id":"https://openalex.org/C2775945657","wikidata":"https://www.wikidata.org/wiki/Q381442","display_name":"Structuring","level":2,"score":0.5044999718666077},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.4918000102043152},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.4681999981403351},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.4404999911785126},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4404999911785126},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.39899998903274536},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.3977999985218048},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.39629998803138733},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.38609999418258667},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.38449999690055847},{"id":"https://openalex.org/C4554734","wikidata":"https://www.wikidata.org/wiki/Q593744","display_name":"Knowledge base","level":2,"score":0.37770000100135803},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.3767000138759613},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.3603000044822693},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.34790000319480896},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.3328999876976013},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.3183000087738037},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2890999913215637},{"id":"https://openalex.org/C25810664","wikidata":"https://www.wikidata.org/wiki/Q44325","display_name":"Ontology","level":2,"score":0.2761000096797943},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.27160000801086426},{"id":"https://openalex.org/C197914299","wikidata":"https://www.wikidata.org/wiki/Q18650","display_name":"Semantic memory","level":3,"score":0.2705000042915344},{"id":"https://openalex.org/C96711827","wikidata":"https://www.wikidata.org/wiki/Q17012245","display_name":"Entity linking","level":3,"score":0.2653999924659729},{"id":"https://openalex.org/C159423971","wikidata":"https://www.wikidata.org/wiki/Q177251","display_name":"Associative property","level":2,"score":0.25760000944137573},{"id":"https://openalex.org/C41009113","wikidata":"https://www.wikidata.org/wiki/Q54871","display_name":"SPARQL","level":4,"score":0.257099986076355},{"id":"https://openalex.org/C85407183","wikidata":"https://www.wikidata.org/wiki/Q1045785","display_name":"Semantic network","level":2,"score":0.25529998540878296},{"id":"https://openalex.org/C72434380","wikidata":"https://www.wikidata.org/wiki/Q230930","display_name":"State space","level":2,"score":0.2502000033855438}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2602.15156","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.15156","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.2602.15156","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.15156","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/4","display_name":"Quality Education","score":0.4183759391307831}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Language":[0],"models":[1],"are":[2],"increasingly":[3],"used":[4],"to":[5,49,141],"reason":[6,50],"over":[7,64],"content":[8],"they":[9],"were":[10],"not":[11],"trained":[12],"on,":[13],"such":[14],"as":[15,122,227],"new":[16,99],"documents,":[17],"evolving":[18],"knowledge,":[19],"and":[20,37,93,109,130,146,172,205,218,237],"user-specific":[21],"data.":[22],"A":[23],"common":[24],"approach":[25],"is":[26,244],"retrieval-augmented":[27],"generation":[28],"(RAG),":[29],"which":[30,116],"stores":[31],"verbatim":[32,168],"documents":[33,121,169],"externally":[34],"(as":[35],"chunks)":[36],"retrieves":[38,173],"only":[39,161],"a":[40,81,158],"relevant":[41],"subset":[42],"at":[43,223,240,246],"inference":[44,152,177],"time":[45,225],"for":[46,138],"an":[47,102,128,139],"LLM":[48,140],"over.":[51],"However,":[52],"this":[53,118],"results":[54,214],"in":[55],"inefficient":[56],"usage":[57],"of":[58,133,221],"test-time":[59],"compute":[60],"(LLM":[61],"repeatedly":[62],"reasons":[63],"the":[65,88,143,155,163,167,174,185,230],"same":[66],"documents);":[67],"moreover,":[68],"chunk":[69],"retrieval":[70],"can":[71],"inject":[72],"irrelevant":[73],"context":[74],"that":[75,107,216],"increases":[76],"unsupported":[77,207],"generation.":[78],"We":[79,113],"propose":[80],"human-like":[82],"non-parametric":[83],"continual":[84],"learning":[85,94],"framework,":[86],"where":[87],"base":[89],"model":[90],"remains":[91],"fixed,":[92],"occurs":[95],"by":[96,119,229],"integrating":[97],"each":[98],"experience":[100],"into":[101],"external":[103],"semantic":[104],"memory":[105],"state":[106],"accumulates":[108],"consolidates":[110],"itself":[111],"continually.":[112],"present":[114],"Panini,":[115],"realizes":[117],"representing":[120],"Generative":[123],"Semantic":[124],"Workspaces":[125],"(GSW)":[126],"--":[127,226,233],"entity-":[129],"event-aware":[131],"network":[132],"question-answer":[134],"(QA)":[135],"pairs,":[136],"sufficient":[137],"reconstruct":[142],"experienced":[144],"situations":[145],"mine":[147],"latent":[148],"knowledge":[149],"via":[150],"reasoning-grounded":[151],"chains":[153],"on":[154,209],"network.":[156],"Given":[157],"query,":[159],"Panini":[160,183],"traverses":[162],"continually-updated":[164],"GSW":[165,231],"(not":[166],"or":[170],"chunks),":[171],"most":[175],"likely":[176],"chains.":[178],"Across":[179],"six":[180],"QA":[181],"benchmarks,":[182],"achieves":[184],"highest":[186],"average":[187],"performance,":[188],"5%-7%":[189],"higher":[190],"than":[191],"other":[192],"competitive":[193],"baselines,":[194],"while":[195],"using":[196],"2-30x":[197],"fewer":[198],"answer-context":[199],"tokens,":[200],"supports":[201],"fully":[202],"open-source":[203],"pipelines,":[204],"reduces":[206],"answers":[208],"curated":[210],"unanswerable":[211],"queries.":[212],"The":[213],"show":[215],"efficient":[217],"accurate":[219],"structuring":[220],"experiences":[222],"write":[224],"achieved":[228],"framework":[232],"yields":[234],"both":[235],"efficiency":[236],"reliability":[238],"gains":[239],"read":[241],"time.":[242],"Code":[243],"available":[245],"https://github.com/roychowdhuryresearch/gsw-memory.":[247]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-02-19T00:00:00"}