{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,27]],"date-time":"2026-03-27T16:14:02Z","timestamp":1774628042550,"version":"3.50.1"},"reference-count":20,"publisher":"IEEE","license":[{"start":{"date-parts":[[2022,3,14]],"date-time":"2022-03-14T00:00:00Z","timestamp":1647216000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2022,3,14]],"date-time":"2022-03-14T00:00:00Z","timestamp":1647216000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key Research and Development Program of China","doi-asserted-by":"publisher","award":["2018YFB1403400"],"award-info":[{"award-number":["2018YFB1403400"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["61834006"],"award-info":[{"award-number":["61834006"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2022,3,14]]},"DOI":"10.23919\/date54114.2022.9774692","type":"proceedings-article","created":{"date-parts":[[2022,5,19]],"date-time":"2022-05-19T20:35:05Z","timestamp":1652992505000},"page":"700-705","source":"Crossref","is-referenced-by-count":25,"title":["DTQAtten: Leveraging Dynamic Token-based Quantization for Efficient Attention Architecture"],"prefix":"10.23919","author":[{"given":"Tao","family":"Yang","sequence":"first","affiliation":[{"name":"Shanghai Jiao Tong University,Shanghai,China"}]},{"given":"Dongyue","family":"Li","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,Shanghai,China"}]},{"given":"Zhuoran","family":"Song","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,Shanghai,China"}]},{"given":"Yilong","family":"Zhao","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,Shanghai,China"}]},{"given":"Fangxin","family":"Liu","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,Shanghai,China"}]},{"given":"Zongwu","family":"Wang","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,Shanghai,China"}]},{"given":"Zhezhi","family":"He","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,Shanghai,China"}]},{"given":"Li","family":"Jiang","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University,Shanghai,China"}]}],"member":"263","reference":[{"key":"ref10","article-title":"A3: Accelerating attention mechanisms in neural networks with approximation","author":"ham","year":"0","journal-title":"HPCA"},{"key":"ref11","article-title":"Attention is all you need","author":"vaswani","year":"0","journal-title":"NIPS"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2020\/520"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3079856.3080246"},{"key":"ref14","article-title":"Gemmini: Enabling systematic deep-learning architecture evaluation via full-stack integration","author":"gene","year":"0","journal-title":"DAC"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/EMC2-NIPS53020.2019.00016"},{"key":"ref16","article-title":"Practical bayesian optimization of machine learning algorithms","author":"snoek","year":"0","journal-title":"NIPS"},{"key":"ref17","author":"yu","year":"2020","journal-title":"Search what you want Barrier panelty nas for mixed precision quantization"},{"key":"ref18","author":"chelba","year":"2013","journal-title":"One Billion Word Benchmark for Measuring Progress in Statistical Language Modelling"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2016.2616357"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00215"},{"key":"ref3","article-title":"Language models are unsupervised multitask learners","author":"radford","year":"2019","journal-title":"OpenAIRE blog"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i05.6409"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA45697.2020.00086"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA51647.2021.00018"},{"key":"ref7","article-title":"I-bert: Integer-only bert quantization","author":"kim","year":"0","journal-title":"ICML"},{"key":"ref2","author":"devlin","year":"2019","journal-title":"BERT Pre-training of deep bidirectional transformers for language understanding"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2020.2979670"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1145\/3307650.3322214"},{"key":"ref20","article-title":"Cacti 6.0: A tool to model large caches","year":"2009","journal-title":"Bragantia"}],"event":{"name":"2022 Design, Automation & Test in Europe Conference & Exhibition (DATE)","location":"Antwerp, Belgium","start":{"date-parts":[[2022,3,14]]},"end":{"date-parts":[[2022,3,23]]}},"container-title":["2022 Design, Automation & Test in Europe Conference & Exhibition (DATE)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/9774496\/9774497\/09774692.pdf?arnumber=9774692","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,7,11]],"date-time":"2022-07-11T20:06:33Z","timestamp":1657569993000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9774692\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,3,14]]},"references-count":20,"URL":"https:\/\/doi.org\/10.23919\/date54114.2022.9774692","relation":{},"subject":[],"published":{"date-parts":[[2022,3,14]]}}}