{"id":"https://openalex.org/W4411207917","doi":"https://doi.org/10.1109/coolchips65488.2025.11018582","title":"A Lightweight Transformer Model with Dynamic Sparse Mask for Neural Machine Translation","display_name":"A Lightweight Transformer Model with Dynamic Sparse Mask for Neural Machine Translation","publication_year":2025,"publication_date":"2025-04-16","ids":{"openalex":"https://openalex.org/W4411207917","doi":"https://doi.org/10.1109/coolchips65488.2025.11018582"},"language":"en","primary_location":{"id":"doi:10.1109/coolchips65488.2025.11018582","is_oa":false,"landing_page_url":"https://doi.org/10.1109/coolchips65488.2025.11018582","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Symposium on Low-Power and High-Speed Chips and Systems (COOL CHIPS)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Nastaran Asadi","orcid":null},"institutions":[{"id":"https://openalex.org/I75917431","display_name":"Nara Institute of Science and Technology","ror":"https://ror.org/05bhada84","country_code":"JP","type":"education","lineage":["https://openalex.org/I75917431"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Nastaran Asadi","raw_affiliation_strings":["Nara Institute of Science and Technology,Division of Information Science,Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nara Institute of Science and Technology,Division of Information Science,Japan","institution_ids":["https://openalex.org/I75917431"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092921203","display_name":"Babak Golbabaei","orcid":"https://orcid.org/0009-0002-4843-5145"},"institutions":[{"id":"https://openalex.org/I75917431","display_name":"Nara Institute of Science and Technology","ror":"https://ror.org/05bhada84","country_code":"JP","type":"education","lineage":["https://openalex.org/I75917431"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Babak Golbabaei","raw_affiliation_strings":["Nara Institute of Science and Technology,Division of Information Science,Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nara Institute of Science and Technology,Division of Information Science,Japan","institution_ids":["https://openalex.org/I75917431"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051037426","display_name":"Yirong Kan","orcid":"https://orcid.org/0000-0002-4070-0672"},"institutions":[{"id":"https://openalex.org/I75917431","display_name":"Nara Institute of Science and Technology","ror":"https://ror.org/05bhada84","country_code":"JP","type":"education","lineage":["https://openalex.org/I75917431"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Yirong Kan","raw_affiliation_strings":["Nara Institute of Science and Technology,Division of Information Science,Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nara Institute of Science and Technology,Division of Information Science,Japan","institution_ids":["https://openalex.org/I75917431"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072048679","display_name":"Renyuan Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I75917431","display_name":"Nara Institute of Science and Technology","ror":"https://ror.org/05bhada84","country_code":"JP","type":"education","lineage":["https://openalex.org/I75917431"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Renyuan Zhang","raw_affiliation_strings":["Nara Institute of Science and Technology,Division of Information Science,Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nara Institute of Science and Technology,Division of Information Science,Japan","institution_ids":["https://openalex.org/I75917431"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5074853381","display_name":"Yasuhiko Nakashima","orcid":"https://orcid.org/0000-0002-9457-5061"},"institutions":[{"id":"https://openalex.org/I75917431","display_name":"Nara Institute of Science and Technology","ror":"https://ror.org/05bhada84","country_code":"JP","type":"education","lineage":["https://openalex.org/I75917431"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Yasuhiko Nakashima","raw_affiliation_strings":["Nara Institute of Science and Technology,Division of Information Science,Japan"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nara Institute of Science and Technology,Division of Information Science,Japan","institution_ids":["https://openalex.org/I75917431"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":1.7588,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.86941312,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.941100001335144,"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":0.941100001335144,"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/transformer","display_name":"Transformer","score":0.7254554629325867},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7183359265327454},{"id":"https://openalex.org/keywords/machine-translation","display_name":"Machine translation","score":0.6224544048309326},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.43076765537261963},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.210744708776474},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.1948322355747223},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.15196558833122253}],"concepts":[{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.7254554629325867},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7183359265327454},{"id":"https://openalex.org/C203005215","wikidata":"https://www.wikidata.org/wiki/Q79798","display_name":"Machine translation","level":2,"score":0.6224544048309326},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.43076765537261963},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.210744708776474},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.1948322355747223},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.15196558833122253}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/coolchips65488.2025.11018582","is_oa":false,"landing_page_url":"https://doi.org/10.1109/coolchips65488.2025.11018582","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Symposium on Low-Power and High-Speed Chips and Systems (COOL CHIPS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.41999998688697815,"id":"https://metadata.un.org/sdg/7","display_name":"Affordable and clean energy"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W3176018525","https://openalex.org/W2903533908","https://openalex.org/W2903810591","https://openalex.org/W2963499882","https://openalex.org/W3026554633","https://openalex.org/W2888520903","https://openalex.org/W2890256614","https://openalex.org/W3098873988","https://openalex.org/W2949454572","https://openalex.org/W3066373881"],"abstract_inverted_index":{"Neural":[0],"machine":[1],"translation":[2,9,37,42,84],"(NMT)":[3],"models,":[4],"especially":[5],"Transformers,":[6,63],"have":[7],"high":[8,29],"quality":[10,43],"but":[11],"suffer":[12],"from":[13],"scalability":[14],"issues":[15],"due":[16],"to":[17,28,100,118,129,161],"the":[18,22,89,96,107,114,120,130,142],"quadratic":[19],"complexity":[20],"of":[21,122,139],"attention":[23,92],"mechanism.":[24],"This":[25],"drawback":[26],"leads":[27],"memory":[30],"usage":[31],"and":[32,148],"computational":[33,46],"overhead,":[34],"making":[35],"large-scale":[36],"inefficient.":[38],"In":[39],"addition,":[40],"maintaining":[41],"while":[44,82],"reducing":[45,162],"costs":[47],"remains":[48],"a":[49],"key":[50],"challenge.":[51],"To":[52],"alleviate":[53],"these":[54],"issues,":[55],"this":[56],"work":[57],"presents":[58],"dynamic":[59],"sparse":[60],"masks":[61],"for":[62],"which":[64,158],"fine-grained":[65],"control":[66],"sparsity":[67],"based":[68],"on":[69,113,125],"percentile":[70],"thresholds.":[71],"Our":[72],"method":[73],"greatly":[74],"reduces":[75,149],"computation":[76],"amount":[77],"in":[78,141,154,165],"Transformer":[79,132,155],"inference":[80,156],"process":[81],"preserving":[83],"accuracy":[85,103],"by":[86,104],"selectively":[87],"retaining":[88],"most":[90],"significant":[91],"scores.":[93],"Furthermore,":[94],"learning":[95],"curriculum":[97],"is":[98,159],"incorporated":[99],"improve":[101],"model":[102,124,135],"gradually":[105],"structuring":[106],"training":[108],"process.":[109],"Experiments":[110],"are":[111],"conducted":[112],"WMT2014":[115],"EnglishFrench":[116],"dataset":[117],"verify":[119],"effectiveness":[121],"our":[123,134],"NMT":[126],"tasks.":[127],"Compared":[128],"standard":[131],"model,":[133],"achieves":[136],"an":[137],"improvement":[138],"19.38":[140],"Bilingual":[143],"Evaluation":[144],"Understudy":[145],"(BLEU)":[146],"score":[147],"31.4%":[150],"multiply-accumulate":[151],"(MAC)":[152],"operations":[153],"process,":[157],"crucial":[160],"power":[163],"consumption":[164],"hardware.":[166]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}