{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,21]],"date-time":"2026-07-21T12:08:50Z","timestamp":1784635730506,"version":"3.55.0"},"reference-count":139,"publisher":"Association for Computing Machinery (ACM)","issue":"4","license":[{"start":{"date-parts":[[2022,11,21]],"date-time":"2022-11-21T00:00:00Z","timestamp":1668988800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"DOI":"10.13039\/501100001659","name":"German Research Foundation","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"crossref"}]},{"name":"ACCROSS: Approximate Computing aCROss the System Stack","award":["2343\/16-1"],"award-info":[{"award-number":["2343\/16-1"]}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Comput. Surv."],"published-print":{"date-parts":[[2023,4,30]]},"abstract":"<jats:p>Deep Neural Networks (DNNs) are very popular because of their high performance in various cognitive tasks in Machine Learning (ML). Recent advancements in DNNs have brought levels beyond human accuracy in many tasks, but at the cost of high computational complexity. To enable efficient execution of DNN inference, more and more research works, therefore, are exploiting the inherent error resilience of DNNs and employing Approximate Computing (AC) principles to address the elevated energy demands of DNN accelerators. This article provides a comprehensive survey and analysis of hardware approximation techniques for DNN accelerators. First, we analyze the state of the art, and by identifying approximation families, we cluster the respective works with respect to the approximation type. Next, we analyze the complexity of the performed evaluations (with respect to the dataset and DNN size) to assess the efficiency, potential, and limitations of approximate DNN accelerators. Moreover, a broad discussion is provided regarding error metrics that are more suitable for designing approximate units for DNN accelerators as well as accuracy recovery approaches that are tailored to DNN inference. Finally, we present how Approximate Computing for DNN accelerators can go beyond energy efficiency and address reliability and security issues as well.<\/jats:p>","DOI":"10.1145\/3527156","type":"journal-article","created":{"date-parts":[[2022,3,25]],"date-time":"2022-03-25T13:06:20Z","timestamp":1648213580000},"page":"1-36","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":129,"title":["Hardware Approximate Techniques for Deep Neural Network Accelerators: A Survey"],"prefix":"10.1145","volume":"55","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-7184-3740","authenticated-orcid":false,"given":"Giorgos","family":"Armeniakos","sequence":"first","affiliation":[{"name":"National Technical University of Athens, Athens, Greece"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Georgios","family":"Zervakis","sequence":"additional","affiliation":[{"name":"Karlsruhe Institute of Technology, Karlsruhe, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dimitrios","family":"Soudris","sequence":"additional","affiliation":[{"name":"National Technical University of Athens, Athens, Greece"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"J\u00f6rg","family":"Henkel","sequence":"additional","affiliation":[{"name":"Karlsruhe Institute of Technology, Karlsruhe, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2022,11,21]]},"reference":[{"key":"e_1_3_2_2_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISSCC42613.2021.9365791"},{"key":"e_1_3_2_3_2","doi-asserted-by":"crossref","first-page":"92","DOI":"10.1109\/ARITH.2019.00023","volume-title":"2019 IEEE 26th Symposium on Computer Arithmetic (ARITH\u201919)","author":"Agrawal A.","year":"2019","unstructured":"A. Agrawal et\u00a0al. 2019. DLFloat: A 16-b floating point format designed for deep learning training and inference. In 2019 IEEE 26th Symposium on Computer Arithmetic (ARITH\u201919). 92\u201395."},{"key":"e_1_3_2_4_2","first-page":"1","volume-title":"IEEE Symposium in Low-Power and High-Speed Chips","author":"Bahou A. Al","year":"2018","unstructured":"A. Al Bahou, G. Karunaratne, R. Andri, L. Cavigelli, and L. Benini. 2018. XNORBIN: A 95 TOp\/s\/W hardware accelerator for binary convolutional neural networks. In IEEE Symposium in Low-Power and High-Speed Chips. 1\u20133."},{"key":"e_1_3_2_5_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2020.3012753"},{"key":"e_1_3_2_6_2","first-page":"236","volume-title":"Computer Society Annual Symposium on VLSI (ISVLSI\u201916)","author":"Andri Renzo","year":"2016","unstructured":"Renzo Andri, Lukas Cavigelli, Davide Rossi, and L. Benini. 2016. YodaNN: An ultra-low power convolutional neural network accelerator based on binary weights. In Computer Society Annual Symposium on VLSI (ISVLSI\u201916). 236\u2013241."},{"key":"e_1_3_2_7_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2019.2940943"},{"key":"e_1_3_2_8_2","volume-title":"Arm Ethos-N Processors","year":"2020","unstructured":"Arm. 2020. Arm Ethos-N Processors. https:\/\/developer.arm.com\/ip-products\/processors\/machine-learning\/arm-ethos-n."},{"key":"e_1_3_2_9_2","article-title":"Post training 4-bit quantization of convolutional networks for rapid-deployment","volume":"32","author":"Banner Ron","year":"2019","unstructured":"Ron Banner, Yury Nahshan, and Daniel Soudry. 2019. Post training 4-bit quantization of convolutional networks for rapid-deployment. Advances in Neural Information Processing Systems 32 (2019).","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_2_10_2","first-page":"841","volume-title":"IEEE 16th International Symposium on Biomedical Imaging","author":"Barata C.","year":"2019","unstructured":"C. Barata and J. S. Marques. 2019. Deep learning for skin cancer diagnosis with hierarchical architectures. In IEEE 16th International Symposium on Biomedical Imaging. 841\u2013845."},{"key":"e_1_3_2_11_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISQED.2014.6783335"},{"key":"e_1_3_2_12_2","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1995.7.1.108"},{"key":"e_1_3_2_13_2","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2020.3039858"},{"key":"e_1_3_2_14_2","volume-title":"Cerebras Wafer Scale Engine","year":"2021","unstructured":"Cerebras. 2021. Cerebras Wafer Scale Engine. https:\/\/cerebras.net\/."},{"key":"e_1_3_2_15_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE.2018.8342119"},{"key":"e_1_3_2_16_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P18-1008"},{"key":"e_1_3_2_17_2","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2014.58"},{"key":"e_1_3_2_18_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.eng.2020.01.007"},{"key":"e_1_3_2_19_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2016.40"},{"key":"e_1_3_2_20_2","doi-asserted-by":"publisher","DOI":"10.1109\/JETCAS.2019.2910232"},{"key":"e_1_3_2_21_2","first-page":"348","volume-title":"Proc. of Machine Learning and Systems","author":"Choi Jungwook","year":"2019","unstructured":"Jungwook Choi, Swagath Venkataramani, Vijayalakshmi (Viji) Srinivasan, Kailash Gopalakrishnan, Zhuo Wang, and Pierce Chuang. 2019. Accurate and efficient 2-bit quantized neural networks. In Proc. of Machine Learning and Systems, A. Talwalkar, V. Smith, and M. Zaharia (Eds.), Vol. 1. 348\u2013359."},{"key":"e_1_3_2_22_2","article-title":"PACT: Parameterized clipping activation for quantized neural networks","author":"Choi J.","year":"2018","unstructured":"J. Choi, Z. Wang, S. Venkataramani, P. I-Jen Chuang, V. Srinivasan, and K. Gopalakrishnan. 2018. PACT: Parameterized clipping activation for quantized neural networks. ArXiv (2018). http:\/\/arxiv.org\/abs\/1503.02531.","journal-title":"ArXiv"},{"key":"e_1_3_2_23_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2018.2857019"},{"key":"e_1_3_2_24_2","first-page":"3123","volume-title":"Proc. of the 28th Int. Conf. on Neural Information Processing Systems","author":"Courbariaux M.","year":"2015","unstructured":"M. Courbariaux, Y. Bengio, and J. David. 2015. BinaryConnect: Training deep neural networks with binary weights during propagations. In Proc. of the 28th Int. Conf. on Neural Information Processing Systems. 3123\u20133131."},{"key":"e_1_3_2_25_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE48585.2020.9116476"},{"key":"e_1_3_2_26_2","first-page":"365","volume-title":"2021 26th Asia and South Pacific Design Automation Conference (ASP-DAC\u201921)","author":"Parra Cecilia De la","year":"2021","unstructured":"Cecilia De la Parra, Andre Guntoro, and A. Kumar. 2021. Efficient accuracy recovery in approximate neural networks by systematic error modelling. In 2021 26th Asia and South Pacific Design Automation Conference (ASP-DAC\u201921). 365\u2013371."},{"key":"e_1_3_2_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"e_1_3_2_28_2","first-page":"1","article-title":"Model compression and hardware acceleration for neural networks: A comprehensive survey","author":"Deng Lei","year":"2020","unstructured":"Lei Deng, Guoqi Li, Song Han, L. P. Shi, and Yuan Xie. 2020. Model compression and hardware acceleration for neural networks: A comprehensive survey. Proc. IEEE 108 (2020), 1\u201348.","journal-title":"Proc. IEEE"},{"key":"e_1_3_2_29_2","article-title":"Reduced-precision memory value approximation for deep learning","author":"Deng Zhaoxia","year":"2015","unstructured":"Zhaoxia Deng, C. Xu, Qiong Cai, P. Faraboschi, and H. Packard. 2015. Reduced-precision memory value approximation for deep learning. ArXiv. https:\/\/arxiv.org\/abs\/1511.05236.","journal-title":"ArXiv"},{"key":"e_1_3_2_30_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2019.2939429"},{"key":"e_1_3_2_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/VLSIC.2018.8502276"},{"key":"e_1_3_2_32_2","article-title":"Attacking binarized neural networks","author":"Galloway Angus","year":"2018","unstructured":"Angus Galloway, Graham W. Taylor, and Medhat Moussa. 2018. Attacking binarized neural networks. ArXiv (2018). https:\/\/arxiv.org\/abs\/1711.00449.","journal-title":"ArXiv"},{"key":"e_1_3_2_33_2","article-title":"A survey of quantization methods for efficient neural network inference","author":"Gholami Amir","year":"2021","unstructured":"Amir Gholami, Sehoon Kim, Zhen Dong, Zhewei Yao, Michael W. Mahoney, and Kurt Keutzer. 2021. A survey of quantization methods for efficient neural network inference. ArXiv (2021). https:\/\/arxiv.org\/abs\/2103.13630.","journal-title":"ArXiv"},{"key":"e_1_3_2_34_2","doi-asserted-by":"publisher","DOI":"10.5555\/3086952"},{"key":"e_1_3_2_35_2","unstructured":"Priya Goyal et\u00a0al. 2018. Accurate Large Minibatch SGD: Training ImageNet in 1 Hour. arxiv:1706.02677"},{"key":"e_1_3_2_36_2","volume-title":"Intelligence Processing Unit","year":"2020","unstructured":"Graphcore. 2020. Intelligence Processing Unit. https:\/\/www.graphcore.ai\/products\/ipu."},{"key":"e_1_3_2_37_2","volume-title":"Tensor Streaming Processor","year":"2021","unstructured":"Groq. 2021. Tensor Streaming Processor. https:\/\/groq.com\/technology\/."},{"key":"e_1_3_2_38_2","article-title":"Defensive approximation: Enhancing CNNs security through approximate computing","author":"Guesmi Amira","year":"2020","unstructured":"Amira Guesmi et\u00a0al. 2020. Defensive approximation: Enhancing CNNs security through approximate computing. ArXiv (2020). https:\/\/arxiv.org\/abs\/2006.07700.","journal-title":"ArXiv"},{"key":"e_1_3_2_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/ASP-DAC47756.2020.9045176"},{"key":"e_1_3_2_40_2","unstructured":"Suyog Gupta Ankur Agrawal Kailash Gopalakrishnan and Pritish Narayanan. 2015. Deep learning with limited numerical precision. In Proceedings of the 32nd International Conference on International Conference on Machine Learning Volume 37. 1737\u20131746."},{"key":"e_1_3_2_41_2","first-page":"(2018) 5784\u2013578","article-title":"Ristretto: A framework for empirical study of resource-efficient inference in convolutional neural networks","author":"Gysel P.","year":"2018","unstructured":"P. Gysel, J. Pimentel, M. Motamedi, and S. Ghiasi. 2018. Ristretto: A framework for empirical study of resource-efficient inference in convolutional neural networks. IEEE Trans. on Neural Networks and Learning Sys. 29, 11 (2018) 5784\u20135789.","journal-title":"IEEE Trans. on Neural Networks and Learning Sys."},{"key":"e_1_3_2_42_2","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2021.3049299"},{"key":"e_1_3_2_43_2","doi-asserted-by":"publisher","DOI":"10.1109\/ETS.2013.6569370"},{"key":"e_1_3_2_44_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2016.30"},{"key":"e_1_3_2_45_2","article-title":"CANN: Curable approximations for high-performance deep neural network accelerators","author":"Hanif Muhammad Abdullah","year":"2019","unstructured":"Muhammad Abdullah Hanif, Faiq Khalid, and Muhammad Shafique. 2019. CANN: Curable approximations for high-performance deep neural network accelerators In. Design Automation Conference (DAC\u201919).","journal-title":"Design Automation Conference (DAC\u201919)"},{"key":"e_1_3_2_46_2","doi-asserted-by":"publisher","DOI":"10.1109\/MDAT.2021.3069952"},{"key":"e_1_3_2_47_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_3_2_48_2","volume-title":"Proc. of the Int. Symp. on Low Power Electronics and Design","author":"He Xin","year":"2018","unstructured":"Xin He, Liu Ke, Wenyan Lu, Guihai Yan, and Xuan Zhang. 2018. AxTrain: Hardware-oriented neural network training for approximate inference. In Proc. of the Int. Symp. on Low Power Electronics and Design. Article 20, 6 pages."},{"key":"e_1_3_2_49_2","doi-asserted-by":"crossref","unstructured":"Maedeh Hemmat Joshua San Miguel and Azadeh Davoodi. 2020. AirNN: A featherweight framework for dynamic input-dependent approximation of CNNs. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems 40 10 (2021) 2090\u20132103.","DOI":"10.1109\/TCAD.2020.3033750"},{"key":"e_1_3_2_50_2","volume-title":"NIPS Deep Learning and Representation Learning Workshop","author":"Hinton Geoffrey","year":"2015","unstructured":"Geoffrey Hinton, Oriol Vinyals, and Jeffrey Dean. 2015. Distilling the knowledge in a neural network. In NIPS Deep Learning and Representation Learning Workshop."},{"key":"e_1_3_2_51_2","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1997.9.8.1735"},{"key":"e_1_3_2_52_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISSCC.2014.6757323"},{"key":"e_1_3_2_53_2","first-page":"102","article-title":"A multiplication reduction technique with near-zero approximation for embedded learning in IoT devices","author":"Huan Yuxiang","year":"2016","unstructured":"Yuxiang Huan, Yifan Qin, Yantian You, Lirong Zheng, and Zhuo Zou. 2016. A multiplication reduction technique with near-zero approximation for embedded learning in IoT devices. In International System on Chip Conference, 102\u2013107.","journal-title":"International System on Chip Conference"},{"key":"e_1_3_2_54_2","first-page":"4114","volume-title":"International Conference on Neural Information Processing Systems","author":"Hubara Itay","year":"2016","unstructured":"Itay Hubara, Matthieu Courbariaux, Daniel Soudry, Ran El-Yaniv, and Yoshua Bengio. 2016. Binarized neural networks. In International Conference on Neural Information Processing Systems. 4114\u20134122."},{"key":"e_1_3_2_55_2","first-page":"448","volume-title":"Proceedings of the 32nd International Conference on Machine Learning","volume":"37","author":"Ioffe Sergey","year":"2015","unstructured":"Sergey Ioffe and Christian Szegedy. 2015. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In Proceedings of the 32nd International Conference on Machine Learning, Vol. 37. 448\u2013456."},{"key":"e_1_3_2_56_2","first-page":"2704","article-title":"Quantization and training of neural networks for efficient integer-arithmetic-only inference","author":"Jacob Benoit","year":"2018","unstructured":"Benoit Jacob et\u00a0al. 2018. Quantization and training of neural networks for efficient integer-arithmetic-only inference. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. 2704\u20132713.","journal-title":"Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition"},{"key":"e_1_3_2_57_2","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2020.3006451"},{"key":"e_1_3_2_58_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE.2018.8342202"},{"key":"e_1_3_2_59_2","unstructured":"Norman Jouppi et\u00a0al. 2017. In-datacenter performance analysis of a tensor processing unit. SIGARCH Comput. Archit. News 45 2 (jun 2017) 1\u201312."},{"key":"e_1_3_2_60_2","first-page":"4345","volume-title":"Conf. on Computer Vision and Pattern Recognition (CVPR\u201919)","author":"Jung Sangil","year":"2019","unstructured":"Sangil Jung et\u00a0al. 2019. Learning to quantize deep networks by optimizing quantization intervals with task loss. In Conf. on Computer Vision and Pattern Recognition (CVPR\u201919). 4345\u20134354."},{"key":"e_1_3_2_61_2","first-page":"182","volume-title":"Int. Symp. on On-Line Testing and Robust System Design (IOLTS\u201919)","author":"Khalid Faiq","year":"2019","unstructured":"Faiq Khalid et\u00a0al. 2019. QuSecNets: Quantization-based defense mechanism for securing deep neural network against adversarial attacks. In Int. Symp. on On-Line Testing and Robust System Design (IOLTS\u201919). 182\u2013187."},{"key":"e_1_3_2_62_2","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2020.3029235"},{"key":"e_1_3_2_63_2","doi-asserted-by":"publisher","DOI":"10.1109\/TC.2018.2880742"},{"key":"e_1_3_2_64_2","doi-asserted-by":"publisher","DOI":"10.1145\/3352460.3358280"},{"key":"e_1_3_2_65_2","doi-asserted-by":"crossref","first-page":"166","DOI":"10.1145\/3352460.3358280","volume-title":"Proc. of the 52nd Annual IEEE\/ACM Int. Symp. on Microarchitecture","author":"Koppula Skanda","year":"2019","unstructured":"Skanda Koppula et\u00a0al. 2019. EDEN: Enabling energy-efficient, high-performance deep neural network inference using approximate DRAM. In Proc. of the 52nd Annual IEEE\/ACM Int. Symp. on Microarchitecture. 166\u2013181."},{"key":"e_1_3_2_66_2","article-title":"Quantizing deep convolutional networks for efficient inference: A whitepaper","author":"Krishnamoorthi Raghuraman","year":"2018","unstructured":"Raghuraman Krishnamoorthi. 2018. Quantizing deep convolutional networks for efficient inference: A whitepaper. ArXiv (6 2018). http:\/\/arxiv.org\/abs\/1806.08342.","journal-title":"ArXiv"},{"key":"e_1_3_2_67_2","article-title":"Learning multiple layers of features from tiny images","author":"Krizhevsky Alex","year":"2012","unstructured":"Alex Krizhevsky. 2012. Learning multiple layers of features from tiny images. University of Toronto.","journal-title":"University of Toronto"},{"key":"e_1_3_2_68_2","doi-asserted-by":"crossref","unstructured":"Cecilia De la Parra. 2020. Knowledge distillation and gradient estimation for active error compensation in approximate neural networks. (2020).","DOI":"10.23919\/DATE51398.2021.9473990"},{"key":"e_1_3_2_69_2","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1989.1.4.541"},{"key":"e_1_3_2_70_2","doi-asserted-by":"publisher","DOI":"10.1109\/5.726791"},{"key":"e_1_3_2_71_2","doi-asserted-by":"publisher","DOI":"10.1109\/TNN.2010.2049580"},{"key":"e_1_3_2_72_2","article-title":"Pruning and quantization for deep neural network acceleration: A survey","author":"Liang Tailin","year":"2021","unstructured":"Tailin Liang, John Glossner, Lei Wang, and Shaobo Shi. 2021. Pruning and quantization for deep neural network acceleration: A survey. ArXiv (2021). https:\/\/arxiv.org\/abs\/2101.09671.","journal-title":"ArXiv"},{"key":"e_1_3_2_73_2","doi-asserted-by":"publisher","DOI":"10.1109\/HOTCHIPS.2019.8875654"},{"key":"e_1_3_2_74_2","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2018.2858826"},{"key":"e_1_3_2_75_2","article-title":"Neural networks with few multiplications","author":"Lin Zhouhan","year":"2016","unstructured":"Zhouhan Lin, Matthieu Courbariaux, Roland Memisevic, and Yoshua Bengio. 2016. Neural networks with few multiplications. ArXiv (2016). https:\/\/arxiv.org\/abs\/1510.03009.","journal-title":"ArXiv"},{"key":"e_1_3_2_76_2","doi-asserted-by":"publisher","DOI":"10.1109\/TEC.1962.5219391"},{"key":"e_1_3_2_77_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE.2019.8714880"},{"key":"e_1_3_2_78_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.fcij.2017.12.001"},{"key":"e_1_3_2_79_2","doi-asserted-by":"publisher","DOI":"10.23919\/DATE.2017.7926993"},{"key":"e_1_3_2_80_2","article-title":"Design of power-efficient approximate multipliers for approximate artificial neural networks","author":"Mrazek Vojtech","year":"2016","unstructured":"Vojtech Mrazek, Syed Shakib Sarwar, Lukas Sekanina, Zdenek Vasicek, and K. Roy. 2016. Design of power-efficient approximate multipliers for approximate artificial neural networks. In Int. Conf. on Computer-Aided Design (ICCAD\u201916).","journal-title":"Int. Conf. on Computer-Aided Design (ICCAD\u201916)"},{"key":"e_1_3_2_81_2","doi-asserted-by":"publisher","DOI":"10.1109\/JETCAS.2020.3032495"},{"key":"e_1_3_2_82_2","article-title":"ALWANN: Automatic layer-wise approximation of deep neural network accelerators without retraining","author":"Mrazek Vojtech","year":"2019","unstructured":"Vojtech Mrazek, Zdenek Vasicek, Lukas Sekanina, Muhammad Abdullah Hanif, and Muhammad Shafique. 2019. ALWANN: Automatic layer-wise approximation of deep neural network accelerators without retraining. In Int. Conference on Computer-Aided Design (ICCAD\u201919).","journal-title":"Int. Conference on Computer-Aided Design (ICCAD\u201919)"},{"key":"e_1_3_2_83_2","first-page":"47","volume-title":"The Security of Machine Learning Systems","author":"Mu\u00f1oz-Gonz\u00e1lez Luis","year":"2019","unstructured":"Luis Mu\u00f1oz-Gonz\u00e1lez and Emil C. Lupu. 2019. The Security of Machine Learning Systems. 47\u201379."},{"key":"e_1_3_2_84_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCIC.2016.7919546"},{"key":"e_1_3_2_85_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW.2017.172"},{"key":"e_1_3_2_86_2","unstructured":"Yuval Netzer Tao Wang Adam Coates Alessandro Bissacco Bo Wu and Andrew Ng. 2011. Reading digits in natural images with unsupervised feature learning. In NIPS Workshop on Deep Learning and Unsupervised Feature Learning ."},{"key":"e_1_3_2_87_2","volume-title":"A100 Tensor Core GPU Architecture","year":"2020","unstructured":"NVIDIA. 2020. A100 Tensor Core GPU Architecture. https:\/\/www.nvidia.com\/content\/dam\/en-zz\/Solutions\/Data-Center\/nvidia-ampere-architecture-whitepaper.pdf."},{"key":"e_1_3_2_88_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2018.2878168"},{"key":"e_1_3_2_89_2","unstructured":"Rasmus Berg Palm. 2012. Prediction as a candidate for learning deep hierarchical models of data. Technical University of Denmark DTU Informatics."},{"key":"e_1_3_2_90_2","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2019.2919463"},{"key":"e_1_3_2_91_2","doi-asserted-by":"publisher","DOI":"10.1145\/3140659.3080254"},{"key":"e_1_3_2_92_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISSCC42613.2021.9365928"},{"key":"e_1_3_2_93_2","first-page":"370","article-title":"A resource-efficient multiplierless systolic array architecture for convolutions in deep networks","volume":"67","author":"Parmar Yashrajsinh","year":"2020","unstructured":"Yashrajsinh Parmar and K. Sridharan. 2020. A resource-efficient multiplierless systolic array architecture for convolutions in deep networks. IEEE Trans. Circuits Syst., II, Exp. Briefs 67, 2 (Feb. 2020), 370\u2013374.","journal-title":"IEEE Trans. Circuits Syst., II, Exp. Briefs"},{"key":"e_1_3_2_94_2","first-page":"1","article-title":"A two-stage operand trimming approximate logarithmic multiplier","author":"Pilipovi\u0107 Ratko","year":"2021","unstructured":"Ratko Pilipovi\u0107, Patricio Buli\u0107, and Uro\u0161 Lotri\u010d. 2021. A two-stage operand trimming approximate logarithmic multiplier. IEEE Transactions on Circuits and Systems I: Regular Papers 68, 6 (2021), 1\u201311.","journal-title":"IEEE Transactions on Circuits and Systems I: Regular Papers"},{"key":"e_1_3_2_95_2","doi-asserted-by":"publisher","DOI":"10.1109\/FPL.2019.00063"},{"key":"e_1_3_2_96_2","first-page":"26","article-title":"Going deeper with embedded FPGA platform for convolutional neural network","author":"Qiu Jiantao","year":"2016","unstructured":"Jiantao Qiu et\u00a0al. 2016. Going deeper with embedded FPGA platform for convolutional neural network. International Symposium on Field-Programmable Gate Arrays (FPGA\u201916). 26\u201335.","journal-title":"International Symposium on Field-Programmable Gate Arrays (FPGA\u201916)"},{"key":"e_1_3_2_97_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2021.3124762"},{"key":"e_1_3_2_98_2","first-page":"525","volume-title":"Computer Vision (ECCV\u201916)","author":"Rastegari Mohammad","year":"2016","unstructured":"Mohammad Rastegari, Vicente Ordonez, Joseph Redmon, and Ali Farhadi. 2016. XNOR-Net: ImageNet classification using binary convolutional neural networks. In Computer Vision (ECCV\u201916), Bastian Leibe, Jiri Matas, Nicu Sebe, and Max Welling (Eds.). Springer International Publishing, Cham, 525\u2013542."},{"key":"e_1_3_2_99_2","volume-title":"Approximate Circuits: Methodologies and CAD (1st ed.)","author":"Reda Sherief","year":"2018","unstructured":"Sherief Reda and Muhammad Shafique. 2018. Approximate Circuits: Methodologies and CAD (1st ed.). Springer Publishing Company, Incorporated."},{"key":"e_1_3_2_100_2","article-title":"A comprehensive survey of neural architecture search: Challenges and solutions","author":"Ren Pengzhen","year":"2021","unstructured":"Pengzhen Ren et\u00a0al. 2021. A comprehensive survey of neural architecture search: Challenges and solutions. ArXiv (2021). https:\/\/arxiv.org\/abs\/2006.02903.","journal-title":"ArXiv"},{"key":"e_1_3_2_101_2","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2020.3008256"},{"key":"e_1_3_2_102_2","first-page":"57","volume-title":"Annual International Symposium on Computer Architecture (ISCA\u201918)","author":"Riera Marc","year":"2018","unstructured":"Marc Riera, Jose-Maria Arnau, and Antonio Gonzalez. 2018. Computation reuse in DNNs by exploiting input similarity. In Annual International Symposium on Computer Architecture (ISCA\u201918). 57\u201368."},{"key":"e_1_3_2_103_2","doi-asserted-by":"publisher","DOI":"10.1145\/3316781.3317784"},{"key":"e_1_3_2_104_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2018.2857262"},{"key":"e_1_3_2_105_2","volume-title":"Design, Automation Test in Europe Conf. Exhibition (DATE\u201921)","author":"Salamin Sami","year":"2021","unstructured":"Sami Salamin, Georgios Zervakis, Ourania Spantidi, Iraklis Anagnostopoulos, J\u00f6rg Henkel, and H. Amrouch. 2021. Reliability-aware quantization for anti-aging NPUs. In Design, Automation Test in Europe Conf. Exhibition (DATE\u201921)."},{"key":"e_1_3_2_106_2","doi-asserted-by":"publisher","DOI":"10.1145\/3097264"},{"key":"e_1_3_2_107_2","volume-title":"pytorchcv  \\( \\cdot \\)  PyPI","author":"Semery Oleg","year":"2021","unstructured":"Oleg Semery. 2021. pytorchcv \\( \\cdot \\) PyPI. https:\/\/pypi.org\/project\/pytorchcv\/."},{"key":"e_1_3_2_108_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2016.12"},{"key":"e_1_3_2_109_2","doi-asserted-by":"publisher","DOI":"10.1145\/2744769.2744778"},{"key":"e_1_3_2_110_2","doi-asserted-by":"publisher","DOI":"10.1109\/MDAT.2020.2971217"},{"key":"e_1_3_2_111_2","doi-asserted-by":"publisher","DOI":"10.1109\/LGRS.2010.2052782"},{"key":"e_1_3_2_112_2","doi-asserted-by":"publisher","DOI":"10.1145\/3195970.3196072"},{"key":"e_1_3_2_113_2","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2018.00069"},{"key":"e_1_3_2_114_2","unstructured":"David R. So Chen Liang and Quoc V. Le. 2019. The Evolved Transformer. arxiv:1901.11117"},{"key":"e_1_3_2_115_2","first-page":"1","volume-title":"Int. Conf. Artificial Intelligence Circuits and Systems","author":"Soliman Taha","year":"2021","unstructured":"Taha Soliman, Cecilia De La Parra, Andre Guntoro, and Norbert Wehn. 2021. Adaptable approximation based on bit decomposition for deep neural network accelerators. In Int. Conf. Artificial Intelligence Circuits and Systems. 1\u20134."},{"issue":"1","key":"e_1_3_2_116_2","first-page":"1929","article-title":"Dropout: A simple way to prevent neural networks from overfitting","volume":"15","author":"Srivastava Nitish","year":"2014","unstructured":"Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. 2014. Dropout: A simple way to prevent neural networks from overfitting. Journal of Machine Learning Research 15, 1 (2014), 1929\u20131958.","journal-title":"Journal of Machine Learning Research"},{"key":"e_1_3_2_117_2","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2017.2761740"},{"key":"e_1_3_2_118_2","unstructured":"Florian Tambon et\u00a0al. 2021. How to Certify Machine Learning Based Safety-critical Systems? A Systematic Literature Review. arxiv:2107.12045 [cs.LG]"},{"key":"e_1_3_2_119_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2020.3019460"},{"key":"e_1_3_2_120_2","first-page":"870","article-title":"Approximated prediction strategy for reducing power consumption of convolutional neural network processor","author":"Ujiie T.","year":"2016","unstructured":"T. Ujiie, M. Hiromoto, and T. Sato. 2016. Approximated prediction strategy for reducing power consumption of convolutional neural network processor. In Conf. on Comp. Vision and Pattern Recog. Workshops (CVPRW\u201916). 870\u2013876.","journal-title":"Conf. on Comp. Vision and Pattern Recog. Workshops (CVPRW\u201916)"},{"key":"e_1_3_2_121_2","first-page":"307","volume-title":"2018 28th Int. Conf. on Field Programmable Logic and Applications (FPL\u201918)","author":"Umuroglu Y.","year":"2018","unstructured":"Y. Umuroglu, L. Rasnayake, and M. Sj\u00e4lander. 2018. BISMO: A scalable bit-serial matrix multiplication overlay for reconfigurable computing. In 2018 28th Int. Conf. on Field Programmable Logic and Applications (FPL\u201918). 307\u20133077."},{"key":"e_1_3_2_122_2","first-page":"96","article-title":"Automated circuit approximation method driven by data distribution","author":"Vasicek Zdenek","year":"2019","unstructured":"Zdenek Vasicek, Vojtech Mrazek, and Lukas Sekanina. 2019. Automated circuit approximation method driven by data distribution. InDesign, Automation and Test in Europe Conference Exhibition (DATE\u201919). 96\u2013101.","journal-title":"Design, Automation and Test in Europe Conference Exhibition (DATE\u201919)"},{"key":"e_1_3_2_123_2","doi-asserted-by":"publisher","DOI":"10.1109\/TEVC.2014.2336175"},{"key":"e_1_3_2_124_2","doi-asserted-by":"crossref","unstructured":"F. Vaverka V. Mrazek Z. Vasicek and L. Sekanina. 2020. TFApprox: Towards a Fast Emulation of DNN Approximate Hardware Accelerators on GPU. 294\u2013297.","DOI":"10.23919\/DATE48585.2020.9116299"},{"key":"e_1_3_2_125_2","first-page":"27","volume-title":"Int. Symp. on Low Power Electronics and Design (ISLPED\u201914)","author":"Venkataramani Swagath","year":"2014","unstructured":"Swagath Venkataramani, Ashish Ranjan, K. Roy, and A. Raghunathan. 2014. AxNN: Energy-efficient neuromorphic systems using approximate computing. In Int. Symp. on Low Power Electronics and Design (ISLPED\u201914). 27\u201332."},{"key":"e_1_3_2_126_2","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2020.3029453"},{"key":"e_1_3_2_127_2","first-page":"1","article-title":"Efficient hardware acceleration of CNNs using logarithmic data representation with arbitrary log-base","author":"Vogel Sebastian","year":"2018","unstructured":"Sebastian Vogel, Mengyu Liang, Andre Guntoro, Walter Stechele, and Gerd Ascheid. 2018. Efficient hardware acceleration of CNNs using logarithmic data representation with arbitrary log-base. In IEEE\/ACM International Conference on Computer-Aided Design, Digest of Technical Papers (ICCAD\u201918). 1\u20138.","journal-title":"IEEE\/ACM International Conference on Computer-Aided Design, Digest of Technical Papers (ICCAD\u201918)"},{"key":"e_1_3_2_128_2","doi-asserted-by":"crossref","unstructured":"Sebastian Vogel Jannik Springer Andre Guntoro and Gerd Ascheid. 2019. Self-supervised quantization of pre-trained neural networks for multiplierless acceleration. In 2019 Design Automation & Test in Europe Conference & Exhibition (DATE) . 1094\u20131099.","DOI":"10.23919\/DATE.2019.8714901"},{"key":"e_1_3_2_129_2","volume-title":"NeurIPS","author":"Wang Naigang","year":"2018","unstructured":"Naigang Wang, Jungwook Choi, D. Brand, Chia-Yu Chen, and K. Gopalakrishnan. 2018. Training deep neural networks with 8-bit floating point numbers. In NeurIPS."},{"key":"e_1_3_2_130_2","doi-asserted-by":"publisher","DOI":"10.1109\/HOTCHIPS.2019.8875671"},{"key":"e_1_3_2_131_2","doi-asserted-by":"publisher","DOI":"10.1109\/ACSSC.2017.8335698"},{"key":"e_1_3_2_132_2","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2020.2981395"},{"key":"e_1_3_2_133_2","doi-asserted-by":"publisher","DOI":"10.1145\/3394885.3431632"},{"key":"e_1_3_2_134_2","doi-asserted-by":"publisher","DOI":"10.1109\/DAC18074.2021.9586092"},{"key":"e_1_3_2_135_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2016.2535398"},{"key":"e_1_3_2_136_2","first-page":"373","volume-title":"Computer Vision (ECCV\u201918)","author":"Zhang Dongqing","year":"2018","unstructured":"Dongqing Zhang, Jiaolong Yang, Dongqiangzi Ye, and Gang Hua. 2018. LQ-nets: Learned quantization for highly accurate and compact deep neural networks. In Computer Vision (ECCV\u201918), Vittorio Ferrari, Martial Hebert, Cristian Sminchisescu, and Yair Weiss (Eds.). 373\u2013390."},{"key":"e_1_3_2_137_2","doi-asserted-by":"publisher","DOI":"10.7873\/DATE.2015.0618"},{"key":"e_1_3_2_138_2","article-title":"Incremental network quantization: Towards lossless CNNs with low-precision weights","author":"Zhou Aojun","year":"2017","unstructured":"Aojun Zhou, Anbang Yao, Yiwen Guo, L. Xu, and Y. Chen. 2017. Incremental network quantization: Towards lossless CNNs with low-precision weights. ArXiv (2017). https:\/\/arxiv.org\/abs\/1702.03044.","journal-title":"ArXiv"},{"key":"e_1_3_2_139_2","unstructured":"Shuchang Zhou Yuxin Wu Zekun Ni Xinyu Zhou He Wen and Yuheng Zou. 2018. DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients. arxiv:1606.06160 [cs.NE]"},{"key":"e_1_3_2_140_2","volume-title":"5th International Conference on Learning Representations (ICLR\u201917)","author":"Zhu Chenzhuo","year":"2017","unstructured":"Chenzhuo Zhu, Song Han, Huizi Mao, and William J. Dally. 2017. Trained ternary quantization. In 5th International Conference on Learning Representations (ICLR\u201917)."}],"container-title":["ACM Computing Surveys"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3527156","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3527156","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,17]],"date-time":"2025-06-17T17:51:00Z","timestamp":1750182660000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3527156"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,11,21]]},"references-count":139,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2023,4,30]]}},"alternative-id":["10.1145\/3527156"],"URL":"https:\/\/doi.org\/10.1145\/3527156","relation":{},"ISSN":["0360-0300","1557-7341"],"issn-type":[{"value":"0360-0300","type":"print"},{"value":"1557-7341","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,11,21]]},"assertion":[{"value":"2021-06-17","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-03-14","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-11-21","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}