{"id":"https://openalex.org/W7214928843","doi":"https://doi.org/10.18653/2026.emnlp-main.1714","title":"Less Uniform Discrete Diffusion is More Powerful and Scalable","display_name":"Less Uniform Discrete Diffusion is More Powerful and Scalable","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7214928843","doi":"https://doi.org/10.18653/2026.emnlp-main.1714"},"language":"en","primary_location":{"id":"doi:10.18653/2026.emnlp-main.1714","doi":"https://doi.org/10.18653/2026.emnlp-main.1714","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.1714","pdf_url":"https://aclanthology.org/2026.emnlp-main.1714.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://aclanthology.org/2026.emnlp-main.1714.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5153610472","display_name":"Kaibo Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]},{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN","HK"],"is_corresponding":false,"raw_author_name":"Kaibo Wang","raw_affiliation_strings":["Huawei Foundation Model Department","The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Foundation Model Department","institution_ids":["https://openalex.org/I2250955327"]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153598543","display_name":"Ding Ding","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]},{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN","HK"],"is_corresponding":false,"raw_author_name":"Ding Ding","raw_affiliation_strings":["Huawei Foundation Model Department","The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Foundation Model Department","institution_ids":["https://openalex.org/I2250955327"]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153606412","display_name":"Fangyu Ding","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]},{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN","HK"],"is_corresponding":false,"raw_author_name":"Fangyu Ding","raw_affiliation_strings":["Huawei Foundation Model Department","The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Foundation Model Department","institution_ids":["https://openalex.org/I2250955327"]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153610853","display_name":"Zijin Feng","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zijin Feng","raw_affiliation_strings":["Huawei Foundation Model Department"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Foundation Model Department","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153605845","display_name":"Han Shi","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Han Shi","raw_affiliation_strings":["Huawei Foundation Model Department"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Foundation Model Department","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153608948","display_name":"Haili Bai","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Haoli Bai","raw_affiliation_strings":["Huawei Foundation Model Department"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Foundation Model Department","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153639577","display_name":"Jiacheng Sun","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiacheng Sun","raw_affiliation_strings":["Huawei Foundation Model Department"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Foundation Model Department","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5153582295","display_name":"Yang Xiang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yang Xiang","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"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":0.65432246,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"39089","last_page":"39109"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.6352999806404114,"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":0.6352999806404114,"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":0.2957000136375427,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.03519999980926514,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/discrete-diffusion","display_name":"discrete diffusion","score":0.6269999742507935},{"id":"https://openalex.org/keywords/model-scaling","display_name":"model scaling","score":0.49799999594688416},{"id":"https://openalex.org/keywords/complex-reasoning","display_name":"complex reasoning","score":0.44600000977516174}],"concepts":[{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.7307000160217285},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6715999841690063},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6617000102996826},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.6097000241279602},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.5777999758720398},{"id":"https://openalex.org/C2776650193","wikidata":"https://www.wikidata.org/wiki/Q264661","display_name":"Obstacle","level":2,"score":0.5257999897003174},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.5210000276565552},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4731000065803528},{"id":"https://openalex.org/C2781140086","wikidata":"https://www.wikidata.org/wiki/Q557945","display_name":"Confusion","level":2,"score":0.41100001335144043},{"id":"https://openalex.org/C5297727","wikidata":"https://www.wikidata.org/wiki/Q786970","display_name":"Autocorrelation","level":2,"score":0.4049000144004822},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.3741999864578247},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.326200008392334},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.3240000009536743},{"id":"https://openalex.org/C55128770","wikidata":"https://www.wikidata.org/wiki/Q5275440","display_name":"Diffusion map","level":4,"score":0.31369999051094055},{"id":"https://openalex.org/C121864883","wikidata":"https://www.wikidata.org/wiki/Q677916","display_name":"Statistical physics","level":1,"score":0.3100000023841858},{"id":"https://openalex.org/C55689738","wikidata":"https://www.wikidata.org/wiki/Q15963867","display_name":"Discrete time and continuous time","level":2,"score":0.29989999532699585},{"id":"https://openalex.org/C141123601","wikidata":"https://www.wikidata.org/wiki/Q6935072","display_name":"Multiscale modeling","level":2,"score":0.2854999899864197},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2831000089645386},{"id":"https://openalex.org/C68710425","wikidata":"https://www.wikidata.org/wiki/Q5275442","display_name":"Diffusion process","level":3,"score":0.27239999175071716},{"id":"https://openalex.org/C184720557","wikidata":"https://www.wikidata.org/wiki/Q7825049","display_name":"Topology (electrical circuits)","level":2,"score":0.26919999718666077},{"id":"https://openalex.org/C571446","wikidata":"https://www.wikidata.org/wiki/Q6510183","display_name":"Diffusion equation","level":3,"score":0.2662000060081482},{"id":"https://openalex.org/C198082294","wikidata":"https://www.wikidata.org/wiki/Q3399648","display_name":"Position (finance)","level":2,"score":0.2565999925136566},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.2506999969482422}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.18653/2026.emnlp-main.1714","doi":"https://doi.org/10.18653/2026.emnlp-main.1714","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.1714","pdf_url":"https://aclanthology.org/2026.emnlp-main.1714.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2609.35817","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35817","pdf_url":"https://arxiv.org/pdf/2609.35817","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,"listed_in":[],"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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.35817","doi":"https://doi.org/10.48550/arxiv.2609.35817","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.35817","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,"listed_in":[],"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"}],"best_oa_location":{"id":"doi:10.18653/2026.emnlp-main.1714","doi":"https://doi.org/10.18653/2026.emnlp-main.1714","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.1714","pdf_url":"https://aclanthology.org/2026.emnlp-main.1714.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214928843.pdf","grobid_xml":"https://content.openalex.org/works/W7214928843.grobid-xml"},"referenced_works_count":34,"referenced_works":["https://openalex.org/W2606964149","https://openalex.org/W2794325560","https://openalex.org/W2890894339","https://openalex.org/W2936262349","https://openalex.org/W2946609015","https://openalex.org/W3083410900","https://openalex.org/W3177813494","https://openalex.org/W3195140297","https://openalex.org/W3210904070","https://openalex.org/W4322718191","https://openalex.org/W4387964173","https://openalex.org/W4388718089","https://openalex.org/W4388926373","https://openalex.org/W4407679583","https://openalex.org/W4414585681","https://openalex.org/W4414858786","https://openalex.org/W4415336572","https://openalex.org/W4416048624","https://openalex.org/W4416051103","https://openalex.org/W4416113503","https://openalex.org/W4416543926","https://openalex.org/W4417297081","https://openalex.org/W4417465430","https://openalex.org/W4417529490","https://openalex.org/W7084115179","https://openalex.org/W7101591989","https://openalex.org/W7111177649","https://openalex.org/W7117673088","https://openalex.org/W7125406378","https://openalex.org/W7128767804","https://openalex.org/W7130703811","https://openalex.org/W7134817167","https://openalex.org/W7140667739","https://openalex.org/W7154221019"],"related_works":["https://openalex.org/W7223246099","https://openalex.org/W7171555131","https://openalex.org/W7127296995","https://openalex.org/W7165160316","https://openalex.org/W7160170245","https://openalex.org/W7164897106","https://openalex.org/W7101591989","https://openalex.org/W7162128388","https://openalex.org/W7129372038","https://openalex.org/W7131431151"],"abstract_inverted_index":{"Although":[0],"uniform":[1,46],"diffusion":[2,9,117],"language":[3],"models":[4],"(UDLMs)":[5],"represent":[6],"a":[7,37,44,89,97,111],"promising":[8],"paradigm,":[10],"scaling":[11],"them":[12],"remains":[13,129],"challenging.We":[14],"identify":[15],"the":[16,54,60,121],"core":[17],"obstacle":[18],"as":[19],"an":[20],"overuniform":[21],"training":[22],"objective":[23],"and":[24,57,83,110],"condition-target":[25],"confusion":[26],"during":[27],"sampling.To":[28],"address":[29],"these,":[30],"we":[31,41],"propose":[32],"Less":[33],"Uniform":[34],"Diffusion":[35],"(LUDI),":[36],"novel":[38],"UDLM":[39,98],"framework.Specifically,":[40],"(i)":[42],"introduce":[43],"less":[45],"loss":[47],"that":[48,66,78,120],"directs":[49],"each":[50],"reverse":[51],"transition":[52],"toward":[53],"clean":[55],"token,":[56],"(ii)":[58],"equip":[59],"model":[61,92],"with":[62,115],"per-token":[63],"time":[64],"embeddings":[65],"supply":[67],"token-level":[68],"corruption":[69],"hints,":[70],"enabling":[71],"confidence-based":[72],"few-step":[73,85],"sampling.Experiments":[74],"across":[75],"scales":[76],"show":[77],"LUDI":[79],"yields":[80],"cleaner":[81],"supervision":[82],"improves":[84],"generation.We":[86],"further":[87],"continue-train":[88],"7B":[90],"autoregressive":[91],"into":[93],"LUDI-7B,":[94],"resulting":[95],"in":[96],"capable":[99],"of":[100,124],"complex":[101,127],"reasoning.It":[102],"achieves":[103],"3":[104],"token-per-step":[105],"speedup":[106],"over":[107],"AR":[108],"decoding":[109],"competitive":[112],"performance":[113],"compared":[114],"masked":[116],"baselines,":[118],"revealing":[119],"full":[122],"potential":[123],"UDLMs":[125],"for":[126],"generation":[128],"to":[130],"be":[131],"unlocked.":[132]},"counts_by_year":[],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-10-01T00:00:00"}