{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T17:44:41Z","timestamp":1777657481940,"version":"3.51.4"},"reference-count":65,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"12","license":[{"start":{"date-parts":[[2023,12,1]],"date-time":"2023-12-01T00:00:00Z","timestamp":1701388800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2023,12,1]],"date-time":"2023-12-01T00:00:00Z","timestamp":1701388800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2023,12,1]],"date-time":"2023-12-01T00:00:00Z","timestamp":1701388800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Research Grants Council of the Hong Kong Special Administrative Region, China","award":["HKU RFS2122-7S04"],"award-info":[{"award-number":["HKU RFS2122-7S04"]}]},{"DOI":"10.13039\/501100021171","name":"Basic and Applied Basic Research Foundation of Guangdong Province","doi-asserted-by":"publisher","award":["2019B1515130003"],"award-info":[{"award-number":["2019B1515130003"]}],"id":[{"id":"10.13039\/501100021171","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Hong Kong Research Grants Council","award":["17212423"],"award-info":[{"award-number":["17212423"]}]},{"name":"Shenzhen Science and Technology Program","award":["JCYJ20200109141414409"],"award-info":[{"award-number":["JCYJ20200109141414409"]}]},{"name":"Hong Kong Research Grant Council - Early Career Scheme","award":["27209621"],"award-info":[{"award-number":["27209621"]}]},{"name":"General Research Fund Scheme","award":["17202422"],"award-info":[{"award-number":["17202422"]}]},{"name":"RGC matching fund scheme"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Pattern Anal. Mach. Intell."],"published-print":{"date-parts":[[2023,12]]},"DOI":"10.1109\/tpami.2023.3319045","type":"journal-article","created":{"date-parts":[[2023,9,25]],"date-time":"2023-09-25T18:19:16Z","timestamp":1695665956000},"page":"15964-15978","source":"Crossref","is-referenced-by-count":7,"title":["Vertical Layering of Quantized Neural Networks for Heterogeneous Inference"],"prefix":"10.1109","volume":"45","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-2921-9132","authenticated-orcid":false,"given":"Hai","family":"Wu","sequence":"first","affiliation":[{"name":"Department of Electrical and Electronic Engineering, The University of Hong Kong, Pok Fu Lam, Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-6161-005X","authenticated-orcid":false,"given":"Ruifei","family":"He","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering, The University of Hong Kong, Pok Fu Lam, Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-6721-2468","authenticated-orcid":false,"given":"Haoru","family":"Tan","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering, The University of Hong Kong, Pok Fu Lam, Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4285-1626","authenticated-orcid":false,"given":"Xiaojuan","family":"Qi","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering, The University of Hong Kong, Pok Fu Lam, Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8773-4629","authenticated-orcid":false,"given":"Kaibin","family":"Huang","sequence":"additional","affiliation":[{"name":"Department of Electrical and Electronic Engineering, The University of Hong Kong, Pok Fu Lam, Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref57","first-page":"2111","article-title":"Direct quantization for training highly accurate low bit-width deep neural networks","author":"hoang","year":"2021","journal-title":"Proc Int Conf Artif Intell"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0816-y"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00448"},{"key":"ref15","first-page":"9847","article-title":"Towards accurate post-training network quantization via bit-split and stitching","author":"wang","year":"2020","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref59","article-title":"SGDR: Stochastic gradient descent with warm restarts","author":"loshchilov","year":"2017","journal-title":"Proc Int Conf Learn Represent"},{"key":"ref14","first-page":"4466","article-title":"Accurate post training quantization with small calibration sets","author":"hubara","year":"2021","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref58","first-page":"87.1","article-title":"Wide residual networks","author":"zagoruyko","year":"2016","journal-title":"Proc Brit Mach Vis Conf"},{"key":"ref53","article-title":"Fighting quantization bias with bias","author":"finkelstein","year":"2019","journal-title":"Proc IEEE\/CVF Conf Comput Vis Pattern Recog Workshop"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00141"},{"key":"ref11","article-title":"Learning multiple layers of features from tiny images","author":"krizhevsky","year":"2009"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00530"},{"key":"ref10","first-page":"1097","article-title":"ImageNet classification with deep convolutional neural networks","author":"krizhevsky","year":"2012","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00529"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1007\/s10994-021-06053-z"},{"key":"ref16","first-page":"7950","article-title":"Post training 4-bit quantization of convolutional networks for rapid-deployment","author":"banner","year":"2019","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref19","first-page":"3123","article-title":"BinaryConnect: Training deep neural networks with binary weights during propagations","author":"courbariaux","year":"2015","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58536-5_5"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2021.3088904"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v35i8.16839"},{"key":"ref46","first-page":"12169","article-title":"In defense of the unitary scalarization for deep multi-task learning","year":"2022","journal-title":"Proc Adv Neural Inform Process Syst"},{"key":"ref45","first-page":"313","article-title":"Mutiple-gradient descent algorithm for multiobjective optimization","author":"d\u00e9sid\u00e9ri","year":"2012","journal-title":"Proc Eur Congr Comput Methods Appl Sci Eng"},{"key":"ref48","article-title":"Apprentice: Using knowledge distillation techniques to improve low-precision network accuracy","author":"mishra","year":"2017","journal-title":"arXiv 1711 05852"},{"key":"ref47","article-title":"Distilling the knowledge in a neural network","author":"hinton","year":"2014","journal-title":"Proc Adv Neural Inf Process Syst Workshop"},{"key":"ref42","volume":"164","author":"hwang","year":"2012","journal-title":"Multiple Objective Decision Making Methods and Application - A State of the Art Survey"},{"key":"ref41","volume":"12","author":"miettinen","year":"2012","journal-title":"Nonlinear Multiobjective Optimization"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1137\/15M1016424"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1007\/s001860000043"},{"key":"ref49","article-title":"QKD: Quantization-aware knowledge distillation","author":"kim","year":"2019","journal-title":"arXiv 1911 12491"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00495"},{"key":"ref7","article-title":"PACT: Parameterized clipping activation for quantized neural networks","author":"choi","year":"2018","journal-title":"arXiv 1805 06085"},{"key":"ref9","article-title":"Learned step size quantization","author":"esser","year":"2020","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref4","article-title":"Dorefa-Net: Training low bitwidth convolutional neural networks with low bitwidth gradients","author":"zhou","year":"2016","journal-title":"arXiv 1606 06160 [cs]"},{"key":"ref3","first-page":"6869","article-title":"Quantized neural networks: Training neural networks with low precision weights and activations","volume":"18","author":"hubara","year":"2017","journal-title":"J Mach Learn Res"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01237-3_23"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01225-0_36"},{"key":"ref40","first-page":"525","article-title":"Multi-task learning as multi-objective optimization","author":"sener","year":"2018","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref35","article-title":"Single path one-shot neural architecture search with uniform sampling","author":"guo","year":"2019","journal-title":"arXiv 1904 00420"},{"key":"ref34","first-page":"4095","article-title":"Efficient neural architecture search via parameters sharing","author":"pham","year":"2018","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref37","article-title":"Once for all: Train one network and specialize it for efficient deployment","author":"cai","year":"2020","journal-title":"Proc Int Conf Learn Represent"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.01385"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00919"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01261-8_25"},{"key":"ref33","article-title":"Slimmable neural networks","author":"yu","year":"2019","journal-title":"Proc Int Conf Learn Represent"},{"key":"ref32","first-page":"1886","article-title":"Channel gating neural networks","author":"hua","year":"2019","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref2","article-title":"Compressing deep convolutional networks using vector quantization","author":"gong","year":"2014","journal-title":"arXiv 1412 6115"},{"key":"ref1","article-title":"Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding","author":"han","year":"2016","journal-title":"Proc Int Conf Learn Represent"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00222"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00215"},{"key":"ref24","article-title":"Trained ternary quantization","author":"zhu","year":"2017","journal-title":"Proc Int Conf Learn Represent"},{"key":"ref23","first-page":"525","article-title":"XOR-Net: ImageNet classification using binary convolutional neural networks","author":"rastegari","year":"2016","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00038"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00881"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-019-01227-8"},{"key":"ref64","article-title":"MQBench: Towards reproducible and deployable model quantization benchmark","author":"li","year":"2021"},{"key":"ref63","article-title":"Differentiable model compression via pseudo quantization noise","author":"d\u00e9fossez","year":"2021","journal-title":"arXiv 2104 09987"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.574"},{"key":"ref21","article-title":"Training binary neural networks with real-to-binary convolutions","author":"martinez","year":"2020","journal-title":"Proc Int Conf Learn Represent"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1109\/SiPS50750.2020.9195245"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2021.3117837"},{"key":"ref27","first-page":"11875","article-title":"HAWQ-V3: Dyadic neural network quantization","author":"yao","year":"2021","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref29","first-page":"527","article-title":"Adaptive neural networks for efficient inference","author":"bolukbasi","year":"2017","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00474"},{"key":"ref62","first-page":"8026","article-title":"PyTorch: An imperative style, high-performance deep learning library","author":"paszke","year":"2019","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref61","article-title":"Very deep convolutional networks for large-scale image recognition","author":"simonyan","year":"2015","journal-title":"Proc Int Conf Learn Representations"}],"container-title":["IEEE Transactions on Pattern Analysis and Machine Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/34\/10308548\/10262344.pdf?arnumber=10262344","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,11,27]],"date-time":"2023-11-27T19:52:58Z","timestamp":1701114778000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10262344\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,12]]},"references-count":65,"journal-issue":{"issue":"12"},"URL":"https:\/\/doi.org\/10.1109\/tpami.2023.3319045","relation":{},"ISSN":["0162-8828","2160-9292","1939-3539"],"issn-type":[{"value":"0162-8828","type":"print"},{"value":"2160-9292","type":"electronic"},{"value":"1939-3539","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,12]]}}}