{"id":"https://openalex.org/W7140667739","doi":"https://doi.org/10.48550/arxiv.2603.23507","title":"Beyond Masks: Efficient, Flexible Diffusion Language Models via Deletion-Insertion Processes","display_name":"Beyond Masks: Efficient, Flexible Diffusion Language Models via Deletion-Insertion Processes","publication_year":2026,"publication_date":"2026-03-04","ids":{"openalex":"https://openalex.org/W7140667739","doi":"https://doi.org/10.48550/arxiv.2603.23507"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.23507","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.23507","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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.2603.23507","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5130633941","display_name":"Fangyu Ding","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ding, Fangyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130657797","display_name":"Ding Ding","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ding, Ding","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129930510","display_name":"Sijin Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Sijin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130713084","display_name":"Kaibo Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Kaibo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130646085","display_name":"Peng Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Peng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130673546","display_name":"Zijin Feng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Feng, Zijin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130682116","display_name":"Haoli Bai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bai, Haoli","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130649685","display_name":"Kai Han","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han, Kai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130688460","display_name":"Youliang Yan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yan, Youliang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130700983","display_name":"Binhang Yuan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan, Binhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130715150","display_name":"Jiacheng Sun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Jiacheng","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":0.42640000581741333,"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":0.42640000581741333,"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":0.13230000436306,"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/T13910","display_name":"Computational and Text Analysis Methods","score":0.061900001019239426,"subfield":{"id":"https://openalex.org/subfields/3300","display_name":"General Social Sciences"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.7900999784469604},{"id":"https://openalex.org/keywords/flexibility","display_name":"Flexibility (engineering)","score":0.6531999707221985},{"id":"https://openalex.org/keywords/masking","display_name":"Masking (illustration)","score":0.5720000267028809},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5497999787330627},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.475600004196167},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4634000062942505},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.3853999972343445},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.3743000030517578}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7922000288963318},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7900999784469604},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.6531999707221985},{"id":"https://openalex.org/C2777402240","wikidata":"https://www.wikidata.org/wiki/Q6783436","display_name":"Masking (illustration)","level":2,"score":0.5720000267028809},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5497999787330627},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.475600004196167},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4634000062942505},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.39820000529289246},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.39089998602867126},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.3853999972343445},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.3743000030517578},{"id":"https://openalex.org/C37404715","wikidata":"https://www.wikidata.org/wiki/Q380679","display_name":"Dynamic programming","level":2,"score":0.323199987411499},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.31790000200271606},{"id":"https://openalex.org/C179603123","wikidata":"https://www.wikidata.org/wiki/Q1941921","display_name":"Modeling language","level":3,"score":0.3102000057697296},{"id":"https://openalex.org/C2780385302","wikidata":"https://www.wikidata.org/wiki/Q367158","display_name":"Protocol (science)","level":3,"score":0.2856000065803528},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2825999855995178},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.27970001101493835},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.2791999876499176},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.2741999924182892},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.2718000113964081},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.25999999046325684},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.2547000050544739},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.25049999356269836}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.23507","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.23507","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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.2603.23507","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.23507","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"While":[0],"Masked":[1],"Diffusion":[2,36],"Language":[3],"Models":[4],"(MDLMs)":[5],"relying":[6],"on":[7,78],"token":[8,43,122,136],"masking":[9,28,53],"and":[10,21,45,54,63,86,108,139,164,176,187],"unmasking":[11,55],"have":[12],"shown":[13],"promise":[14],"in":[15,57,74,90,180],"language":[16,37],"modeling,":[17],"their":[18],"computational":[19,72],"efficiency":[20,65],"generation":[22,115],"flexibility":[23,97],"remain":[24],"constrained":[25],"by":[26,66],"the":[27,52,76,84,168],"paradigm.":[29],"In":[30],"this":[31],"paper,":[32],"we":[33,127,151],"propose":[34],"Deletion-Insertion":[35],"models":[38],"(DID)":[39],"that":[40,119,132],"rigorously":[41],"formulate":[42],"deletion":[44],"insertion":[46,118,137],"as":[47],"discrete":[48],"diffusion":[49],"processes,":[50],"replacing":[51],"processes":[56],"current":[58],"MDLMs.":[59],"DID":[60,94,171],"improves":[61],"training":[62,142],"inference":[64],"eliminating":[67],"two":[68],"major":[69],"sources":[70],"of":[71,170,174,182],"overhead":[73],"MDLMs:":[75],"computations":[77],"non-informative":[79],"1)":[80,99],"tokens":[81,88],"inherent":[82],"to":[83,117,135],"paradigm,":[85],"2)":[87,109],"introduced":[89],"variable-length":[91,102,165],"settings.":[92],"Furthermore,":[93],"offers":[95],"greater":[96],"by:":[98],"natively":[100],"supporting":[101],"sequences":[103],"without":[104,190],"requiring":[105],"fixed-length":[106],"padding,":[107],"an":[110],"intrinsic":[111],"self-correction":[112],"mechanism":[113],"during":[114],"due":[116],"dynamically":[120],"adjusts":[121],"positions.":[123],"To":[124],"train":[125],"DID,":[126],"design":[128],"a":[129,155],"score-based":[130],"approach":[131],"assigns":[133],"scores":[134],"operations":[138],"derive":[140],"appropriate":[141],"objectives.":[143],"The":[144],"objectives":[145],"involve":[146],"subsequence":[147],"counting":[148],"problems,":[149],"which":[150],"efficiently":[152],"solve":[153],"via":[154],"parallelized":[156],"dynamic":[157],"programming":[158],"algorithm.":[159],"Our":[160],"experiments":[161],"across":[162],"fixed":[163],"settings":[166],"demonstrate":[167],"advantage":[169],"over":[172],"baselines":[173],"MDLMs":[175],"existing":[177],"insertion-based":[178],"LMs,":[179],"terms":[181],"modeling":[183],"performance,":[184],"sampling":[185],"quality,":[186],"training/inference":[188],"speed,":[189],"any":[191],"hyperparameter":[192],"tuning.":[193]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-03-27T00:00:00"}