{"id":"https://openalex.org/W4383501753","doi":"https://doi.org/10.1109/aicas57966.2023.10168658","title":"In-memory Activation Compression for GPT Training","display_name":"In-memory Activation Compression for GPT Training","publication_year":2023,"publication_date":"2023-06-11","ids":{"openalex":"https://openalex.org/W4383501753","doi":"https://doi.org/10.1109/aicas57966.2023.10168658"},"language":"en","primary_location":{"id":"doi:10.1109/aicas57966.2023.10168658","is_oa":false,"landing_page_url":"https://doi.org/10.1109/aicas57966.2023.10168658","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2023 IEEE 5th International Conference on Artificial Intelligence Circuits and Systems (AICAS)","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":"https://openalex.org/A5101408508","display_name":"Seungyong Lee","orcid":"https://orcid.org/0000-0002-7348-5454"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":true,"raw_author_name":"Seungyong Lee","raw_affiliation_strings":["Seoul National University,ISRC, Electrical and Computer Engineering,Seoul,South Korea","ISRC, Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"],"affiliations":[{"raw_affiliation_string":"Seoul National University,ISRC, Electrical and Computer Engineering,Seoul,South Korea","institution_ids":["https://openalex.org/I139264467"]},{"raw_affiliation_string":"ISRC, Electrical and Computer Engineering, Seoul National University, Seoul, South Korea","institution_ids":["https://openalex.org/I139264467"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104140397","display_name":"Geonu Yun","orcid":null},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Geonu Yun","raw_affiliation_strings":["Seoul National University,ISRC, Electrical and Computer Engineering,Seoul,South Korea","ISRC, Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"],"affiliations":[{"raw_affiliation_string":"Seoul National University,ISRC, Electrical and Computer Engineering,Seoul,South Korea","institution_ids":["https://openalex.org/I139264467"]},{"raw_affiliation_string":"ISRC, Electrical and Computer Engineering, Seoul National University, Seoul, South Korea","institution_ids":["https://openalex.org/I139264467"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5024723558","display_name":"Hyuk\u2010Jae Lee","orcid":"https://orcid.org/0000-0001-6811-9647"},"institutions":[{"id":"https://openalex.org/I139264467","display_name":"Seoul National University","ror":"https://ror.org/04h9pn542","country_code":"KR","type":"education","lineage":["https://openalex.org/I139264467"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Hyuk-Jae Lee","raw_affiliation_strings":["Seoul National University,ISRC, Electrical and Computer Engineering,Seoul,South Korea","ISRC, Electrical and Computer Engineering, Seoul National University, Seoul, South Korea"],"affiliations":[{"raw_affiliation_string":"Seoul National University,ISRC, Electrical and Computer Engineering,Seoul,South Korea","institution_ids":["https://openalex.org/I139264467"]},{"raw_affiliation_string":"ISRC, Electrical and Computer Engineering, Seoul National University, Seoul, South Korea","institution_ids":["https://openalex.org/I139264467"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5101408508"],"corresponding_institution_ids":["https://openalex.org/I139264467"],"apc_list":null,"apc_paid":null,"fwci":0.308,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.48115578,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":94},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9975000023841858,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9975000023841858,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.9926000237464905,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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":0.9843000173568726,"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/computer-science","display_name":"Computer science","score":0.7178020477294922},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.6922552585601807},{"id":"https://openalex.org/keywords/compression-ratio","display_name":"Compression ratio","score":0.6458616852760315},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.6019926071166992},{"id":"https://openalex.org/keywords/data-compression-ratio","display_name":"Data compression ratio","score":0.5855069160461426},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.5289039015769958},{"id":"https://openalex.org/keywords/exponent","display_name":"Exponent","score":0.48604756593704224},{"id":"https://openalex.org/keywords/image-compression","display_name":"Image compression","score":0.35458454489707947},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.3544765114784241},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.20597168803215027}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7178020477294922},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.6922552585601807},{"id":"https://openalex.org/C25797200","wikidata":"https://www.wikidata.org/wiki/Q828137","display_name":"Compression ratio","level":3,"score":0.6458616852760315},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.6019926071166992},{"id":"https://openalex.org/C94835093","wikidata":"https://www.wikidata.org/wiki/Q3113333","display_name":"Data compression ratio","level":5,"score":0.5855069160461426},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5289039015769958},{"id":"https://openalex.org/C2780388253","wikidata":"https://www.wikidata.org/wiki/Q5421508","display_name":"Exponent","level":2,"score":0.48604756593704224},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.35458454489707947},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3544765114784241},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.20597168803215027},{"id":"https://openalex.org/C171146098","wikidata":"https://www.wikidata.org/wiki/Q124192","display_name":"Automotive engineering","level":1,"score":0.0},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0},{"id":"https://openalex.org/C511840579","wikidata":"https://www.wikidata.org/wiki/Q12757","display_name":"Internal combustion engine","level":2,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/aicas57966.2023.10168658","is_oa":false,"landing_page_url":"https://doi.org/10.1109/aicas57966.2023.10168658","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2023 IEEE 5th International Conference on Artificial Intelligence Circuits and Systems (AICAS)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":12,"referenced_works":["https://openalex.org/W1550898840","https://openalex.org/W2052440657","https://openalex.org/W2130252115","https://openalex.org/W2142265837","https://openalex.org/W2486202470","https://openalex.org/W2516109628","https://openalex.org/W2948229371","https://openalex.org/W2962821792","https://openalex.org/W2971115712","https://openalex.org/W4288365403","https://openalex.org/W6679272650","https://openalex.org/W6762407294"],"related_works":["https://openalex.org/W2088378984","https://openalex.org/W2953482628","https://openalex.org/W4220868150","https://openalex.org/W2388481516","https://openalex.org/W2161302774","https://openalex.org/W2386366915","https://openalex.org/W3080614128","https://openalex.org/W2766695209","https://openalex.org/W2022517628","https://openalex.org/W4243608781"],"abstract_inverted_index":{"Recently,":[0],"a":[1,43,51,68,78,121],"large":[2],"number":[3],"of":[4,32,39,62,124,141,156],"parameters":[5],"in":[6,57,67,104],"Transformer-based":[7],"language":[8],"models":[9],"have":[10,25,29],"caused":[11],"memory":[12,63],"short-ages":[13],"during":[14,65],"training.":[15],"Although":[16],"solutions":[17],"such":[18],"as":[19],"mixed":[20],"precision":[21],"and":[22,36,80,89,96,106,133,151],"model":[23,41],"parallelism":[24],"been":[26],"proposed,":[27],"they":[28],"the":[30,40,60,87,94,99,113,127,131,136,149,154,160],"limitation":[31],"inducing":[33],"communication":[34],"overhead":[35],"requiring":[37],"modification":[38],"by":[42,145],"programmer.":[44],"To":[45],"address":[46],"this":[47],"issue,":[48],"we":[49,152],"propose":[50],"scheme":[52],"that":[53],"compresses":[54,81],"activation":[55,75],"data":[56,76],"memory,":[58],"enabling":[59],"reduction":[61],"usage":[64],"training":[66,158],"user-transparent":[69],"manner.":[70],"The":[71,116],"compression":[72,85,92,118,122,139],"algorithm":[73,119],"gathers":[74],"into":[77],"block":[79],"it,":[82],"using":[83],"base-delta":[84],"for":[86,93,126,130,135],"exponent":[88],"bit-plane":[90],"zero":[91],"sign":[95],"mantissa.":[97,137],"Then,":[98],"important":[100],"bits":[101],"are":[102],"arranged":[103],"order,":[105],"LSB":[107],"truncation":[108],"is":[109,143],"applied":[110],"to":[111,148],"fit":[112],"target":[114],"size.":[115],"proposed":[117],"achieves":[120],"ratio":[123,140],"2.09":[125],"sign,":[128],"2.04":[129],"exponent,":[132],"1.21":[134],"A":[138],"3.2":[142],"obtained":[144],"applying":[146],"up":[147],"truncation,":[150],"confirm":[153],"convergence":[155],"GPT-2":[157],"with":[159],"compression.":[161]},"counts_by_year":[{"year":2024,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}
