{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T08:30:48Z","timestamp":1764577848228,"version":"3.46.0"},"reference-count":24,"publisher":"World Scientific Pub Co Pte Ltd","issue":"03","funder":[{"name":"the Key Research and Development Program of Shaanxi","award":["2024GX-ZDCYL-02-15"],"award-info":[{"award-number":["2024GX-ZDCYL-02-15"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J CIRCUIT SYST COMP"],"published-print":{"date-parts":[[2026,2,15]]},"abstract":"<jats:p>The quantization technique of neural networks can achieve a compressed representation of models by reducing weights and activations data bit-width, accelerating the inference process and minimizing models\u2019 consumption of memory occupation and calculation bandwidth. However, existing quantization methods only focus on the change in weight distribution, unable to directly form a feedback relationship between network feature extraction capability and quantization function, which makes the calculation of quantization error and the adjustment of quantization parameters not reasonable, at last leads to unstable convergence and reduced accuracy. In response to these challenges, this paper introduces a quantization-aware training (QAT) approach grounded in the concept of feature loss. Our method places a profound emphasis on quantization\u2019s influence on changes within the feature space, meticulously computing the topological shifts occurring both within and between sample classes. Furthermore, we present a novel framework for the computation of quantization errors, feedback mechanisms and the dynamic updating of quantization parameters. This holistic approach effectively synergizes neural network training with quantization, resulting in significantly enhanced convergence and accuracy, surpassing the performance of existing quantization techniques across different bit-widths. Using the quantization network of ResNet-18 trained on ImageNet100, the Top-1 classification accuracy improves from 71.24% to 72.08% for 3-bit uniform quantization compared to the QAT approach QN. The scheme is verified to reduce the performance gap between quantized and floating-point models.<\/jats:p>","DOI":"10.1142\/s0218126625504092","type":"journal-article","created":{"date-parts":[[2025,7,11]],"date-time":"2025-07-11T08:09:27Z","timestamp":1752221367000},"source":"Crossref","is-referenced-by-count":0,"title":["Bridging the Gap in Neural Network Quantization with Feature Correlation Coding"],"prefix":"10.1142","volume":"35","author":[{"given":"Yueli","family":"Ding","sequence":"first","affiliation":[{"name":"Department of Computer Science and Technology, The Key Laboratory of Smart Human\u2013Computer Interaction and Wearable Technology of Shaanxi Province, Xidian University, Xi\u2019an 710071, P.\u00a0R.\u00a0China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-0772-9339","authenticated-orcid":false,"given":"Linwei","family":"Hu","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Technology, The Key Laboratory of Smart Human\u2013Computer Interaction and Wearable Technology of Shaanxi Province, Xidian University, Xi\u2019an 710071, P.\u00a0R.\u00a0China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Fei","family":"Cheng","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Technology, The Key Laboratory of Smart Human\u2013Computer Interaction and Wearable Technology of Shaanxi Province, Xidian University, Xi\u2019an 710071, P.\u00a0R.\u00a0China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Quan","family":"Wang","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Technology, The Key Laboratory of Smart Human\u2013Computer Interaction and Wearable Technology of Shaanxi Province, Xidian University, Xi\u2019an 710071, P.\u00a0R.\u00a0China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ziheng","family":"Zhu","sequence":"additional","affiliation":[{"name":"Xi\u2019An Aeronautics Computing Technology Research Institute, Xi\u2019an 710119, P.\u00a0R.\u00a0China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Pengfei","family":"Yang","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Technology, The Key Laboratory of Smart Human\u2013Computer Interaction and Wearable Technology of Shaanxi Province, Xidian University, Xi\u2019an 710071, P.\u00a0R.\u00a0China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"219","published-online":{"date-parts":[[2025,8,20]]},"reference":[{"key":"S0218126625504092BIB001","first-page":"1","volume-title":"Proc. Int. Conf. Learning Representations","author":"Chen K.","year":"2021"},{"key":"S0218126625504092BIB003","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v35i4.16462"},{"key":"S0218126625504092BIB004","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP49357.2023.10094626"},{"key":"S0218126625504092BIB005","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58536-5_5"},{"key":"S0218126625504092BIB007","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00286"},{"key":"S0218126625504092BIB008","first-page":"1","volume-title":"Int. Conf. Learning Representations","author":"Bulat A.","year":"2021"},{"key":"S0218126625504092BIB009","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-68238-5_7"},{"key":"S0218126625504092BIB011","doi-asserted-by":"publisher","DOI":"10.1109\/TMM.2021.3124095"},{"key":"S0218126625504092BIB012","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46493-0_32"},{"key":"S0218126625504092BIB014","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00495"},{"key":"S0218126625504092BIB015","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2022.3216776"},{"key":"S0218126625504092BIB016","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2022.3159369"},{"key":"S0218126625504092BIB017","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00240"},{"key":"S0218126625504092BIB018","first-page":"1","volume-title":"Int. Conf. Learning Representations","author":"Alizadeh M.","year":"2020"},{"key":"S0218126625504092BIB021","first-page":"1","volume":"28","author":"Courbariaux M.","year":"2015","journal-title":"Advances in Neural Information Processing Systems"},{"key":"S0218126625504092BIB022","first-page":"1","volume-title":"Int. Conf. Learning Representations","author":"Yin P.","year":"2019"},{"key":"S0218126625504092BIB023","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2020.2976475"},{"key":"S0218126625504092BIB024","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2018.01.010"},{"key":"S0218126625504092BIB025","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2019\/425"},{"key":"S0218126625504092BIB026","first-page":"5700","volume":"44","author":"Young S. I.","year":"2022","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"S0218126625504092BIB027","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2007.1085"},{"key":"S0218126625504092BIB030","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00748"},{"key":"S0218126625504092BIB031","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01237-3_23"},{"volume-title":"Int. Conf. Learning Representations","year":"2017","author":"Zhou A.","key":"S0218126625504092BIB032"}],"container-title":["Journal of Circuits, Systems and Computers"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.worldscientific.com\/doi\/pdf\/10.1142\/S0218126625504092","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T08:28:54Z","timestamp":1764577734000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.worldscientific.com\/doi\/10.1142\/S0218126625504092"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,8,20]]},"references-count":24,"journal-issue":{"issue":"03","published-print":{"date-parts":[[2026,2,15]]}},"alternative-id":["10.1142\/S0218126625504092"],"URL":"https:\/\/doi.org\/10.1142\/s0218126625504092","relation":{},"ISSN":["0218-1266","1793-6454"],"issn-type":[{"type":"print","value":"0218-1266"},{"type":"electronic","value":"1793-6454"}],"subject":[],"published":{"date-parts":[[2025,8,20]]},"article-number":"2550409"}}