{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,19]],"date-time":"2025-12-19T08:47:29Z","timestamp":1766134049213,"version":"3.28.0"},"reference-count":35,"publisher":"IEEE","license":[{"start":{"date-parts":[[2021,6,6]],"date-time":"2021-06-06T00:00:00Z","timestamp":1622937600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2021,6,6]],"date-time":"2021-06-06T00:00:00Z","timestamp":1622937600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2021,6,6]],"date-time":"2021-06-06T00:00:00Z","timestamp":1622937600000},"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":[[2021,6,6]]},"DOI":"10.1109\/icassp39728.2021.9414076","type":"proceedings-article","created":{"date-parts":[[2021,5,13]],"date-time":"2021-05-13T19:53:45Z","timestamp":1620935625000},"page":"7383-7387","source":"Crossref","is-referenced-by-count":11,"title":["Mixed Precision Quantization of Transformer Language Models for Speech Recognition"],"prefix":"10.1109","author":[{"given":"Junhao","family":"Xu","sequence":"first","affiliation":[]},{"given":"Shoukang","family":"Hu","sequence":"additional","affiliation":[]},{"given":"Jianwei","family":"Yu","sequence":"additional","affiliation":[]},{"given":"Xunying","family":"Liu","sequence":"additional","affiliation":[]},{"given":"Helen","family":"Meng","sequence":"additional","affiliation":[]}],"member":"263","reference":[{"key":"ref33","article-title":"Darts: Differentiable architecture search","author":"liu","year":"2018","journal-title":"International Conference on Learning Representations"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01246-5_2"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1145\/1944345.1944349"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP40776.2020.9053483"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/TASL.2010.2064307"},{"key":"ref34","article-title":"Purely sequence-trained neural networks for asr based on lattice-free mmi","author":"galvez","year":"2016","journal-title":"InterSpeech"},{"key":"ref10","article-title":"Experimental determination of precision requirements for back-propagation training of artificial neural networks","author":"asanovic","year":"1991","journal-title":"ICSI"},{"key":"ref11","article-title":"Expectation backpropagation: Parameter-free training of multilayer neural networks with continuous or discrete weights","author":"meir","year":"2014","journal-title":"NIPS"},{"key":"ref12","article-title":"Xnornet: Imagenet classification using binary convolutional neural networks","author":"rastegari","year":"2016","journal-title":"ECCV"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.521"},{"key":"ref14","article-title":"Incremental network quantization: Towards lossless cnns with low-precision weights","author":"zhou","year":"2017","journal-title":"ICLRE"},{"article-title":"Squeezenet: Alexnet-level accuracy with 50x fewer parameters and < 0.5 mb model size","year":"2016","author":"iandola","key":"ref15"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00038"},{"key":"ref17","article-title":"Haq: Hardware-aware automated quantization","author":"liu","year":"2019","journal-title":"CVPR"},{"key":"ref18","first-page":"1137","article-title":"A neural probabilistic language model","volume":"3","author":"bengio","year":"2003","journal-title":"Journal of Machine Learning Research"},{"key":"ref19","article-title":"Quantized neural networks: Training neural networks with low precision weights and activations","author":"bengio","year":"2017","journal-title":"JMLR"},{"article-title":"Distributed optimization and statistical learning via the alternating direction method of multipliers","year":"2011","author":"parikh","key":"ref28"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"article-title":"Ternary weight networks","year":"2016","author":"li","key":"ref27"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1244"},{"key":"ref6","first-page":"5998","article-title":"Attention is all you need","author":"vaswani","year":"2017","journal-title":"NIPS"},{"key":"ref29","article-title":"Alternating multi-bit quantization for recurrent neural networks","author":"xu","year":"2018","journal-title":"International Conference on Learning Representations"},{"article-title":"Layer normalization","year":"2016","author":"ba","key":"ref5"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2019-2225"},{"key":"ref7","first-page":"1243","article-title":"Convolutional sequence to sequence learning","author":"gehring","year":"2017","journal-title":"ICML"},{"key":"ref2","article-title":"A structured self-attentive sentence embedding","author":"lin","year":"2017","journal-title":"Int Conf on Learning Representations (ICLR)"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/ASRU46091.2019.9004025"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1053"},{"key":"ref20","article-title":"Binaryconnect: Training deep neural networks with binary weights during propagations","author":"david","year":"2015","journal-title":"NIPS"},{"key":"ref22","article-title":"Binary neural networks for speech recognition","author":"xiang","year":"2019","journal-title":"Frontiers Inf Technol Electron Eng"},{"article-title":"Hawq-v2: Hessian aware trace-weighted quantization of neural networks","year":"2019","author":"keutzer","key":"ref21"},{"article-title":"Gaussian error linear units (gelus)","year":"2018","author":"hendrycks","key":"ref24"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/TASLP.2020.3015659"},{"key":"ref26","first-page":"525","article-title":"Xnor-net: Imagenet classification using binary convolutional neural networks","author":"rastegari","year":"2016","journal-title":"European Conference on Computer Vision"},{"key":"ref25","doi-asserted-by":"crossref","DOI":"10.1609\/aaai.v32i1.11713","article-title":"Extremely low bit neural network: Squeeze the last bit out with admm","author":"leng","year":"2018","journal-title":"Thirty-Second AAAI Conference on Artificial Intelligence"}],"event":{"name":"ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","start":{"date-parts":[[2021,6,6]]},"location":"Toronto, ON, Canada","end":{"date-parts":[[2021,6,11]]}},"container-title":["ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/9413349\/9413350\/09414076.pdf?arnumber=9414076","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,12,27]],"date-time":"2022-12-27T08:29:42Z","timestamp":1672129782000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9414076\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,6,6]]},"references-count":35,"URL":"https:\/\/doi.org\/10.1109\/icassp39728.2021.9414076","relation":{},"subject":[],"published":{"date-parts":[[2021,6,6]]}}}