{"id":"https://openalex.org/W7134071386","doi":"https://doi.org/10.48550/arxiv.2603.05454","title":"Beyond Scattered Acceptance: Fast and Coherent Inference for DLMs via Longest Stable Prefixes","display_name":"Beyond Scattered Acceptance: Fast and Coherent Inference for DLMs via Longest Stable Prefixes","publication_year":2026,"publication_date":"2026-03-05","ids":{"openalex":"https://openalex.org/W7134071386","doi":"https://doi.org/10.48550/arxiv.2603.05454"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2603.05454","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"type":"article","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128233116","display_name":"Pengxiang Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Pengxiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Tsai, Joey","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tsai, Joey","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128276382","display_name":"Hongwei Xue","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xue, Hongwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128250332","display_name":"Kunyu Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Kunyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128253556","display_name":"Shilin Yan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yan, Shilin","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":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.27615898,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"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/T10181","display_name":"Natural Language Processing Techniques","score":0.3707999885082245,"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":0.3707999885082245,"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":0.26350000500679016,"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":0.07440000027418137,"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/security-token","display_name":"Security token","score":0.7397000193595886},{"id":"https://openalex.org/keywords/prefix","display_name":"Prefix","score":0.7113000154495239},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6395000219345093},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.5092999935150146},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.4221999943256378},{"id":"https://openalex.org/keywords/cyclic-prefix","display_name":"Cyclic prefix","score":0.4185999929904938},{"id":"https://openalex.org/keywords/disjoint-sets","display_name":"Disjoint sets","score":0.37790000438690186},{"id":"https://openalex.org/keywords/boundary","display_name":"Boundary (topology)","score":0.3718999922275543},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.36090001463890076}],"concepts":[{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7397000193595886},{"id":"https://openalex.org/C141603448","wikidata":"https://www.wikidata.org/wiki/Q134830","display_name":"Prefix","level":2,"score":0.7113000154495239},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6395000219345093},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6151999831199646},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.5092999935150146},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5073000192642212},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.4221999943256378},{"id":"https://openalex.org/C2778775284","wikidata":"https://www.wikidata.org/wiki/Q5198224","display_name":"Cyclic prefix","level":4,"score":0.4185999929904938},{"id":"https://openalex.org/C45340560","wikidata":"https://www.wikidata.org/wiki/Q215382","display_name":"Disjoint sets","level":2,"score":0.37790000438690186},{"id":"https://openalex.org/C62354387","wikidata":"https://www.wikidata.org/wiki/Q875399","display_name":"Boundary (topology)","level":2,"score":0.3718999922275543},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.36090001463890076},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.35580000281333923},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.35429999232292175},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.33880001306533813},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.3262999951839447},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.32359999418258667},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.31790000200271606},{"id":"https://openalex.org/C75291252","wikidata":"https://www.wikidata.org/wiki/Q1315756","display_name":"TRACE (psycholinguistics)","level":2,"score":0.3149000108242035},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.2994999885559082},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.2969000041484833},{"id":"https://openalex.org/C184720557","wikidata":"https://www.wikidata.org/wiki/Q7825049","display_name":"Topology (electrical circuits)","level":2,"score":0.29679998755455017},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.2955000102519989},{"id":"https://openalex.org/C159423971","wikidata":"https://www.wikidata.org/wiki/Q177251","display_name":"Associative property","level":2,"score":0.2870999872684479},{"id":"https://openalex.org/C102634674","wikidata":"https://www.wikidata.org/wiki/Q868473","display_name":"Smoothness","level":2,"score":0.28279998898506165},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.2825999855995178},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2824000120162964},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.2685000002384186},{"id":"https://openalex.org/C2778067643","wikidata":"https://www.wikidata.org/wiki/Q166507","display_name":"Interval (graph theory)","level":2,"score":0.2639000117778778},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.258899986743927},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.25839999318122864},{"id":"https://openalex.org/C111696304","wikidata":"https://www.wikidata.org/wiki/Q2303697","display_name":"Sorting","level":2,"score":0.25369998812675476}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2603.05454","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2603.05454","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05454","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":"pmh:doi:10.48550/arxiv.2603.05454","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[{"score":0.7248214483261108,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Diffusion":[0],"Language":[1],"Models":[2],"(DLMs)":[3],"promise":[4],"highly":[5],"parallel":[6],"text":[7],"generation,":[8,175],"yet":[9],"their":[10],"practical":[11,206],"inference":[12,74,163],"speed":[13],"is":[14],"often":[15],"bottlenecked":[16],"by":[17,164],"suboptimal":[18],"decoding":[19],"schedulers.":[20],"Standard":[21],"approaches":[22],"rely":[23],"on":[24,77,155],"'scattered":[25],"acceptance'-committing":[26],"high":[27],"confidence":[28],"tokens":[29],"at":[30],"disjoint":[31],"positions":[32],"throughout":[33],"the":[34,40,49,64,192,197,200],"sequence.":[35],"This":[36,117],"approach":[37],"inadvertently":[38],"fractures":[39],"Key-Value":[41],"(KV)":[42],"cache,":[43],"destroys":[44],"memory":[45],"locality,":[46],"and":[47,72,103,150,157,179,205],"forces":[48],"model":[50],"into":[51,130],"costly,":[52],"repeated":[53],"repairs":[54],"across":[55,168],"unstable":[56],"token":[57,87,147],"boundaries.":[58],"To":[59],"resolve":[60],"this,":[61],"we":[62],"present":[63],"Longest":[65],"Stable":[66],"Prefix":[67],"(LSP)":[68],"scheduler,":[69],"a":[70,90,96,140],"training-free":[71],"model-agnostic":[73],"paradigm":[75],"based":[76],"monolithic":[78],"prefix":[79],"absorption.":[80],"In":[81],"each":[82],"denoising":[83],"step,":[84],"LSP":[85,161,195],"evaluates":[86],"stability":[88],"via":[89],"single":[91],"forward":[92],"pass,":[93],"dynamically":[94],"identifies":[95],"contiguous":[97,132],"left-aligned":[98],"block":[99],"of":[100,203],"stable":[101],"predictions,":[102],"snaps":[104],"its":[105],"boundary":[106],"to":[107,166],"natural":[108],"linguistic":[109],"or":[110,184],"structural":[111],"delimiters":[112],"before":[113],"an":[114],"atomic":[115],"commitment.":[116],"prefix-first":[118],"topology":[119],"yields":[120],"dual":[121],"benefits:":[122],"systemically,":[123],"it":[124,135],"converts":[125],"fragmented":[126],"KV":[127],"cache":[128],"updates":[129],"efficient,":[131],"appends;":[133],"algorithmically,":[134],"preserves":[136],"bidirectional":[137],"lookahead":[138],"over":[139],"geometrically":[141],"shrinking":[142],"active":[143],"suffix,":[144],"drastically":[145],"reducing":[146],"flip":[148],"rates":[149],"denoiser":[151],"calls.":[152],"Extensive":[153],"evaluations":[154],"LLaDA-8B":[156],"Dream-7B":[158],"demonstrate":[159],"that":[160],"accelerates":[162],"up":[165],"3.4x":[167],"rigorous":[169],"benchmarks":[170],"including":[171],"mathematical":[172],"reasoning,":[173],"code":[174],"multilingual":[176],"(CJK)":[177],"tasks,":[178],"creative":[180],"writing":[181],"while":[182],"matching":[183],"slightly":[185],"improving":[186],"output":[187],"quality.":[188],"By":[189],"fundamentally":[190],"restructuring":[191],"commitment":[193],"topology,":[194],"bridges":[196],"gap":[198],"between":[199],"theoretical":[201],"parallelism":[202],"DLMs":[204],"hardware":[207],"efficiency.":[208]},"counts_by_year":[],"updated_date":"2026-07-15T18:14:33.161393","created_date":"2026-03-07T00:00:00"}