{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,21]],"date-time":"2025-11-21T11:33:07Z","timestamp":1763724787784,"version":"3.44.0"},"reference-count":29,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,6,22]],"date-time":"2025-06-22T00:00:00Z","timestamp":1750550400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,6,22]],"date-time":"2025-06-22T00:00:00Z","timestamp":1750550400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,6,22]]},"DOI":"10.1109\/dac63849.2025.11133081","type":"proceedings-article","created":{"date-parts":[[2025,9,15]],"date-time":"2025-09-15T17:35:41Z","timestamp":1757957741000},"page":"1-7","source":"Crossref","is-referenced-by-count":1,"title":["APSQ: Additive Partial Sum Quantization with Algorithm-Hardware Co-Design"],"prefix":"10.1109","author":[{"given":"Yonghao","family":"Tan","sequence":"first","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Pingcheng","family":"Dong","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yongkun","family":"Wu","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yu","family":"Liu","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xuejiao","family":"Liu","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Peng","family":"Luo","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shih-Yang","family":"Liu","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xijie","family":"Huang","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Dong","family":"Zhang","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Luhong","family":"Liang","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kwang-Ting","family":"Cheng","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science Technology,AI Chip Center for Emerging Smart Systems (ACCESS)"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"doi-asserted-by":"publisher","key":"ref1","DOI":"10.48550\/arXiv.2010.11929"},{"doi-asserted-by":"publisher","key":"ref2","DOI":"10.1109\/ICCV51070.2023.01587"},{"doi-asserted-by":"publisher","key":"ref3","DOI":"10.48550\/arXiv.2105.15203"},{"doi-asserted-by":"publisher","key":"ref4","DOI":"10.1109\/TIP.2023.3297408"},{"key":"ref5","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018","journal-title":"arXiv preprint arXiv:1810.04805"},{"key":"ref6","article-title":"Llama 2: Open foundation and fine-tuned chat models","author":"Touvron","year":"2023","journal-title":"arXiv preprint arXiv:2307.09288"},{"doi-asserted-by":"publisher","key":"ref7","DOI":"10.48550\/ARXIV.1706.03762"},{"doi-asserted-by":"publisher","key":"ref8","DOI":"10.1109\/CVPR.2018.00286"},{"key":"ref9","first-page":"5506","article-title":"I-bert: Integer-only bert quantization","volume-title":"International conference on machine learning","author":"Kim"},{"volume-title":"8th International Conference on Learning Representations","author":"Esser","article-title":"Learned step size quantization","key":"ref10"},{"key":"ref11","article-title":"Oscillation-free quantization for low-bit vision transformers","author":"Liu","year":"2023","journal-title":"arXiv preprint arXiv:2302.02210"},{"doi-asserted-by":"publisher","key":"ref12","DOI":"10.1109\/ICCV.2019.00339"},{"doi-asserted-by":"publisher","key":"ref13","DOI":"10.1109\/ICCV48922.2021.00526"},{"doi-asserted-by":"publisher","key":"ref14","DOI":"10.1145\/2654822.2541967"},{"doi-asserted-by":"publisher","key":"ref15","DOI":"10.1109\/jssc.2016.2616357"},{"doi-asserted-by":"publisher","key":"ref16","DOI":"10.1109\/TVLSI.2017.2688340"},{"doi-asserted-by":"publisher","key":"ref17","DOI":"10.1109\/HPCA56546.2023.10071027"},{"doi-asserted-by":"publisher","key":"ref18","DOI":"10.1145\/3466752.3480125"},{"doi-asserted-by":"publisher","key":"ref19","DOI":"10.1109\/TCAD.2023.3294461"},{"doi-asserted-by":"publisher","key":"ref20","DOI":"10.1109\/TCSII.2023.3246562"},{"doi-asserted-by":"publisher","key":"ref21","DOI":"10.1109\/ISSCC.2014.6757323"},{"doi-asserted-by":"publisher","key":"ref22","DOI":"10.1109\/JSSC.2021.3115466"},{"doi-asserted-by":"publisher","key":"ref23","DOI":"10.1109\/JSSC.2022.3141050"},{"key":"ref24","article-title":"Estimating or propagating gradients through stochastic neurons for conditional computation","author":"Bengio","year":"2013","journal-title":"arXiv preprint arXiv:1308.3432"},{"doi-asserted-by":"publisher","key":"ref25","DOI":"10.1145\/3649329.3657314"},{"doi-asserted-by":"publisher","key":"ref26","DOI":"10.18653\/v1\/W18-5446"},{"doi-asserted-by":"publisher","key":"ref27","DOI":"10.1109\/CVPR.2017.544"},{"doi-asserted-by":"publisher","key":"ref28","DOI":"10.18653\/v1\/2024.findings-emnlp.444"},{"key":"ref29","first-page":"8","article-title":"A framework for few-shot language model evaluation","volume":"10","author":"Gao","year":"2021","journal-title":"Version v0. 0.1. Sept"}],"event":{"name":"2025 62nd ACM\/IEEE Design Automation Conference (DAC)","start":{"date-parts":[[2025,6,22]]},"location":"San Francisco, CA, USA","end":{"date-parts":[[2025,6,25]]}},"container-title":["2025 62nd ACM\/IEEE Design Automation Conference (DAC)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11132383\/11132091\/11133081.pdf?arnumber=11133081","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,16]],"date-time":"2025-09-16T05:48:52Z","timestamp":1758001732000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11133081\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,22]]},"references-count":29,"URL":"https:\/\/doi.org\/10.1109\/dac63849.2025.11133081","relation":{},"subject":[],"published":{"date-parts":[[2025,6,22]]}}}