{"id":"https://openalex.org/W7134058767","doi":"https://doi.org/10.48550/arxiv.2603.05172","title":"Trainable Bitwise Soft Quantization for Input Feature Compression","display_name":"Trainable Bitwise Soft Quantization for Input Feature Compression","publication_year":2026,"publication_date":"2026-03-05","ids":{"openalex":"https://openalex.org/W7134058767","doi":"https://doi.org/10.48550/arxiv.2603.05172"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.05172","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05172","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,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2603.05172","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126926779","display_name":"Karsten Schr\u00f6dter","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Schr\u00f6dter, Karsten","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5016217427","display_name":"Jan Stenkamp","orcid":"https://orcid.org/0000-0003-1100-4907"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Stenkamp, Jan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128225349","display_name":"Nina Herrmann","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Herrmann, Nina","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5023532714","display_name":"Fabian Gieseke","orcid":"https://orcid.org/0000-0001-7093-5803"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gieseke, Fabian","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":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":false,"primary_topic":{"id":"https://openalex.org/T10901","display_name":"Advanced Data Compression Techniques","score":0.7150999903678894,"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/T10901","display_name":"Advanced Data Compression Techniques","score":0.7150999903678894,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.14399999380111694,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.01600000075995922,"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/quantization","display_name":"Quantization (signal processing)","score":0.7472000122070312},{"id":"https://openalex.org/keywords/bitwise-operation","display_name":"Bitwise operation","score":0.5817999839782715},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.48429998755455017},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.47110000252723694},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.4632999897003174},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4309000074863434},{"id":"https://openalex.org/keywords/feature-extraction","display_name":"Feature extraction","score":0.40959998965263367},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.37470000982284546}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7710999846458435},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.7472000122070312},{"id":"https://openalex.org/C134765980","wikidata":"https://www.wikidata.org/wiki/Q879126","display_name":"Bitwise operation","level":2,"score":0.5817999839782715},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.48429998755455017},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.47110000252723694},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.4632999897003174},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4528000056743622},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4309000074863434},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.40959998965263367},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.37470000982284546},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.36570000648498535},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.3617999851703644},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.33329999446868896},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.33329999446868896},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.325300008058548},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.32339999079704285},{"id":"https://openalex.org/C134835016","wikidata":"https://www.wikidata.org/wiki/Q690265","display_name":"Lookup table","level":2,"score":0.31949999928474426},{"id":"https://openalex.org/C199833920","wikidata":"https://www.wikidata.org/wiki/Q612536","display_name":"Vector quantization","level":2,"score":0.3003000020980835},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.29820001125335693},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.28790000081062317},{"id":"https://openalex.org/C93996380","wikidata":"https://www.wikidata.org/wiki/Q44127","display_name":"Server","level":2,"score":0.273499995470047},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.26570001244544983},{"id":"https://openalex.org/C94835093","wikidata":"https://www.wikidata.org/wiki/Q3113333","display_name":"Data compression ratio","level":5,"score":0.25459998846054077}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.05172","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05172","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,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.05172","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05172","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,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","score":0.5515756607055664,"display_name":"Affordable and clean energy"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0,145],"growing":[1],"demand":[2],"for":[3,16],"machine":[4],"learning":[5],"applications":[6,44],"in":[7],"the":[8,11,21,51,90,101,111,119,140,218],"context":[9],"of":[10,13,23,93,103,132,142,211,222],"Internet":[12],"Things":[14],"calls":[15],"new":[17],"approaches":[18,56],"to":[19,49,64,74,107,113,125,128,149,185,191,209,224,230],"optimize":[20],"use":[22],"limited":[24],"compute":[25],"and":[26,40],"memory":[27],"resources.":[28,53],"Despite":[29],"significant":[30,234],"progress":[31],"that":[32,88,105,197],"has":[33],"been":[34],"made":[35],"w.r.t.":[36],"reducing":[37],"model":[38],"sizes":[39],"improving":[41],"efficiency,":[42],"many":[43],"still":[45],"require":[46],"remote":[47,65,115],"servers":[48],"provide":[50],"required":[52],"However,":[54],"such":[55],"rely":[57],"on":[58,217],"transmitting":[59],"data":[60,104,143],"from":[61,110,162],"edge":[62],"devices":[63],"servers,":[66],"which":[67],"may":[68],"not":[69],"always":[70],"be":[71,108,126,227],"feasible":[72],"due":[73],"bandwidth,":[75],"latency,":[76],"or":[77],"energy":[78],"constraints.":[79],"We":[80,181],"propose":[81],"a":[82,94,114,129,135,178],"task-specific,":[83],"trainable":[84,156,172],"feature":[85,124],"quantization":[86,157,193,202],"layer":[87,120,146],"compresses":[89],"input":[91,123,232],"features":[92],"neural":[95,179],"network.":[96,180],"This":[97],"can":[98,226],"significantly":[99],"reduce":[100],"amount":[102],"needs":[106],"transferred":[109],"device":[112],"server.":[116],"In":[117,214],"particular,":[118,215],"allows":[121],"each":[122],"quantized":[127,173],"user-defined":[130],"number":[131],"bits,":[133],"enabling":[134,155],"simple":[136],"on-device":[137],"compression":[138,220],"at":[139],"time":[141],"collection.":[144],"is":[147],"designed":[148],"approximate":[150],"step":[151],"functions":[152],"with":[153,177],"sigmoids,":[154,164],"thresholds.":[158],"By":[159],"concatenating":[160],"outputs":[161],"multiple":[163],"introduced":[165],"as":[166,188,190],"bitwise":[167],"soft":[168],"quantization,":[169],"it":[170],"achieves":[171],"values":[174],"when":[175],"integrated":[176],"compare":[182],"our":[183,198],"method":[184],"full-precision":[186,212],"inference":[187],"well":[189],"several":[192],"baselines.":[194],"Experiments":[195],"show":[196],"approach":[199],"outperforms":[200],"standard":[201],"methods,":[203],"while":[204],"maintaining":[205],"accuracy":[206],"levels":[207],"close":[208],"those":[210],"models.":[213],"depending":[216],"dataset,":[219],"factors":[221],"$5\\times$":[223],"$16\\times$":[225],"achieved":[228],"compared":[229],"$32$-bit":[231],"without":[233],"performance":[235],"loss.":[236]},"counts_by_year":[],"updated_date":"2026-08-16T07:02:28.622633","created_date":"2026-03-07T00:00:00"}