{"id":"https://openalex.org/W7166821461","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1819","title":"HTMuon: Improving Muon via Heavy-Tailed Spectral Correction","display_name":"HTMuon: Improving Muon via Heavy-Tailed Spectral Correction","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166821461","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1819"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.1819","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1819","pdf_url":"https://aclanthology.org/2026.findings-acl.1819.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.1819.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5093836321","display_name":"Tianyu Pang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tianyu Pang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128901514","display_name":"Yujie Fang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yujie Fang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004393593","display_name":"Z. Liu","orcid":"https://orcid.org/0009-0004-9638-9803"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zihang Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128902715","display_name":"Shenyang Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shenyang Deng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139765493","display_name":"Lei Hsiung","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lei Hsiung","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100528028","display_name":"Shuhua Yu","orcid":"https://orcid.org/0000-0003-3792-1797"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shuhua Yu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128806929","display_name":"Yaoqing Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yaoqing Yang","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":0.88045302,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"36504","last_page":"36535"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12579","display_name":"Muon and positron interactions and applications","score":0.3626999855041504,"subfield":{"id":"https://openalex.org/subfields/2211","display_name":"Mechanics of Materials"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12579","display_name":"Muon and positron interactions and applications","score":0.3626999855041504,"subfield":{"id":"https://openalex.org/subfields/2211","display_name":"Mechanics of Materials"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10921","display_name":"Neutrino Physics Research","score":0.274399995803833,"subfield":{"id":"https://openalex.org/subfields/3106","display_name":"Nuclear and High Energy Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11044","display_name":"Particle Detector Development and Performance","score":0.09359999746084213,"subfield":{"id":"https://openalex.org/subfields/3106","display_name":"Nuclear and High Energy Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/muon","display_name":"Muon","score":0.4537999927997589},{"id":"https://openalex.org/keywords/detector","display_name":"Detector","score":0.28850001096725464},{"id":"https://openalex.org/keywords/error-detection-and-correction","display_name":"Error detection and correction","score":0.2694000005722046},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.2547000050544739}],"concepts":[{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.671500027179718},{"id":"https://openalex.org/C205334942","wikidata":"https://www.wikidata.org/wiki/Q3151","display_name":"Muon","level":2,"score":0.4537999927997589},{"id":"https://openalex.org/C185544564","wikidata":"https://www.wikidata.org/wiki/Q81197","display_name":"Nuclear physics","level":1,"score":0.3546000123023987},{"id":"https://openalex.org/C109214941","wikidata":"https://www.wikidata.org/wiki/Q18334","display_name":"Particle physics","level":1,"score":0.34360000491142273},{"id":"https://openalex.org/C94915269","wikidata":"https://www.wikidata.org/wiki/Q1834857","display_name":"Detector","level":2,"score":0.28850001096725464},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.28049999475479126},{"id":"https://openalex.org/C103088060","wikidata":"https://www.wikidata.org/wiki/Q1062839","display_name":"Error detection and correction","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.2547000050544739},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.24639999866485596},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.2434999942779541}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.1819","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1819","pdf_url":"https://aclanthology.org/2026.findings-acl.1819.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.1819","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1819","pdf_url":"https://aclanthology.org/2026.findings-acl.1819.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320306084","display_name":"U.S. Department of Energy","ror":"https://ror.org/01bj3aw27"},{"id":"https://openalex.org/F4320310400","display_name":"Dartmouth College","ror":"https://ror.org/049s0rh22"},{"id":"https://openalex.org/F4320332180","display_name":"Defense Advanced Research Projects Agency","ror":"https://ror.org/02caytj08"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166821461.pdf","grobid_xml":"https://content.openalex.org/works/W7166821461.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Muon":[0,89],"has":[1],"recently":[2],"shown":[3],"promising":[4],"results":[5],"in":[6,127],"LLM":[7,64],"training.In":[8],"this":[9],"work,":[10],"we":[11,44],"study":[12],"how":[13],"to":[14,50,104,107,115],"further":[15,109],"improve":[16],"Muon.We":[17,108],"argue":[18],"that":[19,70,112],"Muon's":[20,48],"orthogonalized":[21],"update":[22],"rule":[23],"suppresses":[24],"the":[25,33,39,96,119],"emergence":[26],"of":[27,87,132],"heavy-tailed":[28],"weight":[29,61],"spectra":[30],"and":[31,58,66,78,123],"overemphasizes":[32],"training":[34],"along":[35],"noise-dominated":[36],"directions.Motivated":[37],"by":[38,102],"Heavy-Tailed":[40],"Self-Regularization":[41],"(HT-SR)":[42],"theory,":[43],"propose":[45],"HTMuon.HTMuon":[46],"preserves":[47],"ability":[49],"capture":[51],"parameter":[52],"interdependencies":[53],"while":[54],"producing":[55],"heavier-tailed":[56,60],"updates":[57],"inducing":[59],"spectra.Experiments":[62],"on":[63,85,92,95],"pretraining":[65,94],"image":[67],"classification":[68],"show":[69,111],"HTMuon":[71,99,113,133],"consistently":[72],"improves":[73],"performance":[74],"over":[75],"state-of-the-art":[76],"baselines":[77],"can":[79],"also":[80],"serve":[81],"as":[82],"a":[83],"plug-in":[84],"top":[86],"existing":[88],"variants.For":[90],"example,":[91],"LLaMA":[93],"C4":[97],"dataset,":[98],"reduces":[100],"perplexity":[101],"up":[103],"0.98":[105],"compared":[106],"theoretically":[110],"corresponds":[114],"steepest":[116],"descent":[117],"under":[118],"Schatten-q":[120],"norm":[121],"constraint":[122],"provide":[124],"convergence":[125],"analysis":[126],"smooth":[128],"non-convex":[129],"settings.The":[130],"implementation":[131],"is":[134],"available":[135],"at":[136],"https:":[137],"//github.com/TDCSZ327/HTmuon.":[138]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}
