{"id":"https://openalex.org/W7166846310","doi":"https://doi.org/10.18653/v1/2026.acl-long.2060","title":"Beyond Fully Random Masking: Attention-Guided Denoising and Optimization for Diffusion Language Models","display_name":"Beyond Fully Random Masking: Attention-Guided Denoising and Optimization for Diffusion Language Models","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166846310","doi":"https://doi.org/10.18653/v1/2026.acl-long.2060"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.acl-long.2060","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.2060","pdf_url":"https://aclanthology.org/2026.acl-long.2060.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.acl-long.2060.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139791413","display_name":"Jia Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jia Deng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139849722","display_name":"Junyi Jessy Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Junyi Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139830268","display_name":"Xin Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xin Zhao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139785014","display_name":"Jinpeng Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinpeng Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139849927","display_name":"Hongyu Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hongyu Lu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139737149","display_name":"Ji-Rong Wen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ji-Rong Wen","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":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.87153474,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"44502","last_page":"44514"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.2924000024795532,"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"}},"topics":[{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.2924000024795532,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.14190000295639038,"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":0.08550000190734863,"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/noise-reduction","display_name":"Noise reduction","score":0.3831000030040741},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.3544999957084656},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.3061999976634979},{"id":"https://openalex.org/keywords/diffusion","display_name":"Diffusion","score":0.3043999969959259},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.2567000091075897}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5644999742507935},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4693000018596649},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.44369998574256897},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.3831000030040741},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.3544999957084656},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.31150001287460327},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3061999976634979},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.3043999969959259},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2567000091075897},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.2565000057220459}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.acl-long.2060","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.2060","pdf_url":"https://aclanthology.org/2026.acl-long.2060.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.acl-long.2060","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.2060","pdf_url":"https://aclanthology.org/2026.acl-long.2060.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166846310.pdf","grobid_xml":"https://content.openalex.org/works/W7166846310.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Diffusion":[0],"large":[1],"language":[2],"models":[3,11],"(dLLMs)":[4],"offer":[5],"an":[6,34,68],"efficient":[7],"alternative":[8],"to":[9,48],"autoregressive":[10],"through":[12],"parallel":[13],"decoding,":[14],"yet":[15],"existing":[16],"post-training":[17,115],"methods":[18,116],"largely":[19],"rely":[20],"on":[21,88,101],"random":[22],"masking":[23],"strategies":[24],"that":[25,43,74,107],"overlook":[26],"intrinsic":[27],"token":[28],"dependencies.In":[29],"this":[30],"work,":[31],"we":[32,65],"present":[33],"empirical":[35],"analysis":[36],"of":[37],"attention":[38,89],"in":[39,60],"dLLMs":[40],"and":[41,55,71,78,91,98,103],"show":[42],"tokens":[44,94],"attending":[45],"more":[46],"strongly":[47],"unmasked":[49],"context":[50],"exhibit":[51],"greater":[52],"generation":[53],"stability":[54],"play":[56],"a":[57],"critical":[58],"role":[59],"reasoning.Motivated":[61],"by":[62],"these":[63],"findings,":[64],"propose":[66],"AGDO,":[67],"attention-guided":[69],"denoising":[70,85],"optimization":[72,79],"framework":[73],"aligns":[75],"both":[76],"training":[77],"with":[80],"attention-derived":[81],"dependencies.AGDO":[82],"determines":[83],"the":[84],"order":[86],"based":[87],"structure":[90],"emphasizes":[92],"attentioncritical":[93],"during":[95],"supervised":[96],"fine-tuning":[97],"reinforcement":[99],"learning.Experiments":[100],"mathematical":[102],"coding":[104],"benchmarks":[105],"demonstrate":[106],"AGDO":[108],"consistently":[109],"improves":[110],"reasoning":[111],"performance,":[112],"outperforming":[113],"state-of-theart":[114],"for":[117],"dLLMs.":[118]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}