{"id":"https://openalex.org/W6929607776","doi":"https://doi.org/10.5075/epfl-thesis-9867","title":"Parameter-efficient Methods for Training and Inference of Transformer-based Models","display_name":"Parameter-efficient Methods for Training and Inference of Transformer-based Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W6929607776","doi":"https://doi.org/10.5075/epfl-thesis-9867"},"language":"en","primary_location":{"id":"pmh:oai:infoscience.epfl.ch:20.500.14299/246394","is_oa":true,"landing_page_url":"https://infoscience.epfl.ch/handle/20.500.14299/246394","pdf_url":null,"source":{"id":"https://openalex.org/S4306400487","display_name":"Infoscience (Ecole Polytechnique F\u00e9d\u00e9rale de Lausanne)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"THESIS","raw_type":"doctoral thesis"},"type":"dissertation","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://infoscience.epfl.ch/handle/20.500.14299/246394","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Banaei, Mohammadreza","orcid":null},"institutions":[{"id":"https://openalex.org/I5124864","display_name":"\u00c9cole Polytechnique F\u00e9d\u00e9rale de Lausanne","ror":"https://ror.org/02s376052","country_code":"CH","type":"education","lineage":["https://openalex.org/I2799323385","https://openalex.org/I5124864"]}],"countries":["CH"],"is_corresponding":true,"raw_author_name":"Banaei, Mohammadreza","raw_affiliation_strings":["Ecole polytechnique f\u00e9d\u00e9rale de Lausanne"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Ecole polytechnique f\u00e9d\u00e9rale de Lausanne","institution_ids":["https://openalex.org/I5124864"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I5124864"],"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":true,"primary_topic":{"id":"https://openalex.org/T11901","display_name":"Bayesian Methods and Mixture Models","score":0.017000000923871994,"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/T11901","display_name":"Bayesian Methods and Mixture Models","score":0.017000000923871994,"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/T13141","display_name":"Statistical Methods and Applications","score":0.01269999984651804,"subfield":{"id":"https://openalex.org/subfields/2613","display_name":"Statistics and Probability"},"field":{"id":"https://openalex.org/fields/26","display_name":"Mathematics"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10677","display_name":"Attachment and Relationship Dynamics","score":0.01080000028014183,"subfield":{"id":"https://openalex.org/subfields/3207","display_name":"Social Psychology"},"field":{"id":"https://openalex.org/fields/32","display_name":"Psychology"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/perplexity","display_name":"Perplexity","score":0.7829999923706055},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6704999804496765},{"id":"https://openalex.org/keywords/memory-footprint","display_name":"Memory footprint","score":0.5570999979972839},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.539900004863739},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.5339999794960022},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.492900013923645},{"id":"https://openalex.org/keywords/autoencoder","display_name":"Autoencoder","score":0.4681999981403351},{"id":"https://openalex.org/keywords/rank","display_name":"Rank (graph theory)","score":0.4399000108242035},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.42260000109672546},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.4058000147342682}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8155999779701233},{"id":"https://openalex.org/C100279451","wikidata":"https://www.wikidata.org/wiki/Q372193","display_name":"Perplexity","level":3,"score":0.7829999923706055},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6704999804496765},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5924999713897705},{"id":"https://openalex.org/C74912251","wikidata":"https://www.wikidata.org/wiki/Q6815727","display_name":"Memory footprint","level":2,"score":0.5570999979972839},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5550000071525574},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.539900004863739},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.5339999794960022},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.492900013923645},{"id":"https://openalex.org/C101738243","wikidata":"https://www.wikidata.org/wiki/Q786435","display_name":"Autoencoder","level":3,"score":0.4681999981403351},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.4399000108242035},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.42260000109672546},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.4058000147342682},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.3671000003814697},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.35749998688697815},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.31859999895095825},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.3176000118255615},{"id":"https://openalex.org/C2779439875","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Natural language understanding","level":3,"score":0.3059999942779541},{"id":"https://openalex.org/C132943942","wikidata":"https://www.wikidata.org/wiki/Q2562511","display_name":"Footprint","level":2,"score":0.2863999903202057},{"id":"https://openalex.org/C187834632","wikidata":"https://www.wikidata.org/wiki/Q188804","display_name":"Factorization","level":2,"score":0.28630000352859497},{"id":"https://openalex.org/C2780980858","wikidata":"https://www.wikidata.org/wiki/Q110022","display_name":"Dual (grammatical number)","level":2,"score":0.27900001406669617},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2784000039100647},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.2775000035762787},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.2759999930858612},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2745000123977661},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.27000001072883606},{"id":"https://openalex.org/C124681953","wikidata":"https://www.wikidata.org/wiki/Q339062","display_name":"Decomposition","level":2,"score":0.26969999074935913},{"id":"https://openalex.org/C39927690","wikidata":"https://www.wikidata.org/wiki/Q11197","display_name":"Logarithm","level":2,"score":0.26840001344680786},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.2587999999523163},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.25769999623298645},{"id":"https://openalex.org/C177606310","wikidata":"https://www.wikidata.org/wiki/Q5674297","display_name":"Adaptability","level":2,"score":0.2572999894618988},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.2542000114917755},{"id":"https://openalex.org/C22789450","wikidata":"https://www.wikidata.org/wiki/Q420904","display_name":"Singular value decomposition","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:infoscience.epfl.ch:20.500.14299/246394","is_oa":true,"landing_page_url":"https://infoscience.epfl.ch/handle/20.500.14299/246394","pdf_url":null,"source":{"id":"https://openalex.org/S4306400487","display_name":"Infoscience (Ecole Polytechnique F\u00e9d\u00e9rale de Lausanne)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"THESIS","raw_type":"doctoral thesis"},{"id":"doi:10.5075/epfl-thesis-9867","is_oa":true,"landing_page_url":"https://doi.org/10.5075/epfl-thesis-9867","pdf_url":null,"source":null,"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Dissertation"}],"best_oa_location":{"id":"pmh:oai:infoscience.epfl.ch:20.500.14299/246394","is_oa":true,"landing_page_url":"https://infoscience.epfl.ch/handle/20.500.14299/246394","pdf_url":null,"source":{"id":"https://openalex.org/S4306400487","display_name":"Infoscience (Ecole Polytechnique F\u00e9d\u00e9rale de Lausanne)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"THESIS","raw_type":"doctoral thesis"},"sustainable_development_goals":[{"score":0.7784837484359741,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"advancements":[1],"in":[2,116],"natural":[3],"language":[4,12,61,90,230],"processing,":[5],"driven":[6],"by":[7,50,224],"the":[8,17,55,65,69,81,94,121,143,149,166,173,188,226,233,247],"success":[9],"of":[10,20,60,68,87,97,151,168,190,199,228,270],"large":[11,89],"models,":[13],"have":[14,28],"led":[15],"to":[16,22,79,142,213],"scaling":[18],"up":[19],"models":[21,27,231],"unprecedented":[23],"sizes.":[24],"While":[25],"these":[26,48,200],"achieved":[29],"remarkable":[30],"performance":[31,141,167,211,227],"across":[32,204,250,284],"various":[33,152,205,285],"tasks,":[34,208],"their":[35],"computational":[36,85],"and":[37,57,84,105,207,237,287],"memory":[38,82],"demands":[39],"pose":[40],"significant":[41],"challenges.":[42],"In":[43],"this":[44],"thesis,":[45],"we":[46,72,177],"address":[47,64],"challenges":[49],"introducing":[51],"parameter-efficient":[52,174,181],"methods":[53,184,201],"for":[54,102,120,232],"inference":[56,70,137],"training":[58,175],"phases":[59],"models.":[62,91],"To":[63],"parameter-efficiency":[66],"aspect":[67],"phase,":[71,176],"propose":[73,106],"a":[74,111,158,241,259],"factorization-based":[75],"compression":[76,125,170],"framework":[77,109,126],"designed":[78],"minimize":[80],"footprint":[83],"load":[86],"deploying":[88],"We":[92,146,222],"challenge":[93],"common":[95],"usage":[96],"Singular":[98],"Value":[99],"Decomposition":[100],"(SVD)":[101],"weight":[103,130],"factorization":[104,161],"an":[107],"autoencoder":[108],"with":[110],"custom":[112],"loss":[113],"that":[114,185,239],"results":[115],"better":[117,281],"zero-shot":[118],"perplexity":[119],"compressed":[122],"model.":[123,145],"This":[124],"decomposes":[127],"each":[128],"model's":[129,248],"into":[131],"multiple":[132],"low-rank":[133,261],"weights,":[134],"enabling":[135],"efficient":[136],"while":[138,265],"having":[139,266],"competitive":[140],"original":[144],"also":[147],"demonstrate":[148,238],"effectiveness":[150],"parameter-sharing":[153],"schemes":[154],"as":[155,157],"well":[156],"non-uniform":[159],"(sensitivity-based)":[160],"scheme,":[162],"which":[163],"further":[164],"boost":[165,246],"our":[169,217],"framework.":[171],"Regarding":[172],"present":[178],"two":[179],"novel":[180,260],"fine-tuning":[182],"(PEFT)":[183],"significantly":[186,245],"reduce":[187],"number":[189],"trainable":[191,272],"parameters":[192,273],"without":[193],"sacrificing":[194],"model":[195,288],"performance.":[196],"The":[197],"generalization":[198,249],"is":[202],"assessed":[203],"settings":[206],"demonstrating":[209],"superior":[210],"compared":[212],"baseline":[214],"approaches.":[215],"For":[216],"first":[218],"PEFT":[219,242,255],"method":[220,243,256,263],"contribution,":[221],"start":[223],"studying":[225],"multilingual":[229],"cross-lingual":[234],"reasoning":[235],"task":[236],"incorporating":[240],"can":[244],"different":[251],"tasks/languages.":[252],"Our":[253],"second":[254],"focuses":[257],"on":[258],"adaptation":[262],"that,":[264],"less":[267],"than":[268,282],"1%":[269],"LoRA's":[271],"(Hu":[274],"et":[275],"al.,":[276],"2021),":[277],"performs":[278],"competitively":[279],"or":[280],"LoRA":[283],"benchmarks":[286],"scales.":[289]},"counts_by_year":[],"updated_date":"2026-07-23T05:56:39.545243","created_date":"2025-10-10T00:00:00"}