{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,10]],"date-time":"2026-07-10T15:15:20Z","timestamp":1783696520592,"version":"3.55.0"},"reference-count":41,"publisher":"Elsevier BV","issue":"7","license":[{"start":{"date-parts":[[2026,11,1]],"date-time":"2026-11-01T00:00:00Z","timestamp":1793491200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/tdm\/userlicense\/1.0\/"},{"start":{"date-parts":[[2026,11,1]],"date-time":"2026-11-01T00:00:00Z","timestamp":1793491200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/legal\/tdmrep-license"},{"start":{"date-parts":[[2026,11,1]],"date-time":"2026-11-01T00:00:00Z","timestamp":1793491200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-017"},{"start":{"date-parts":[[2026,11,1]],"date-time":"2026-11-01T00:00:00Z","timestamp":1793491200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,11,1]],"date-time":"2026-11-01T00:00:00Z","timestamp":1793491200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-012"},{"start":{"date-parts":[[2026,11,1]],"date-time":"2026-11-01T00:00:00Z","timestamp":1793491200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,11,1]],"date-time":"2026-11-01T00:00:00Z","timestamp":1793491200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-004"}],"funder":[{"DOI":"10.13039\/501100005230","name":"Natural Science Foundation of Chongqing Municipality","doi-asserted-by":"publisher","award":["CSTB2022NSCQ-MSX1383"],"award-info":[{"award-number":["CSTB2022NSCQ-MSX1383"]}],"id":[{"id":"10.13039\/501100005230","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100007957","name":"Chongqing Municipal Education Commission","doi-asserted-by":"publisher","award":["KJQN202100515"],"award-info":[{"award-number":["KJQN202100515"]}],"id":[{"id":"10.13039\/501100007957","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100007957","name":"Chongqing Municipal Education Commission","doi-asserted-by":"publisher","award":["KJQN202400541"],"award-info":[{"award-number":["KJQN202400541"]}],"id":[{"id":"10.13039\/501100007957","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100010338","name":"Chongqing Normal University","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100010338","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62072097"],"award-info":[{"award-number":["62072097"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["U22A2026"],"award-info":[{"award-number":["U22A2026"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["elsevier.com","sciencedirect.com"],"crossmark-restriction":true},"short-container-title":["Information Processing & Management"],"published-print":{"date-parts":[[2026,11]]},"DOI":"10.1016\/j.ipm.2026.104879","type":"journal-article","created":{"date-parts":[[2026,5,6]],"date-time":"2026-05-06T20:29:53Z","timestamp":1778099393000},"page":"104879","update-policy":"https:\/\/doi.org\/10.1016\/elsevier_cm_policy","source":"Crossref","is-referenced-by-count":0,"special_numbering":"PB","title":["When input perturbation outperforms gradient perturbation: Achieving high-accuracy deep learning under local differential privacy"],"prefix":"10.1016","volume":"63","author":[{"given":"Shunshun","family":"Peng","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jun","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chenxing","family":"Hu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Quanwang","family":"Wu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hongbing","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mengmeng","family":"Yang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4073-0756","authenticated-orcid":false,"given":"Taolin","family":"Guo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"key":"10.1016\/j.ipm.2026.104879_b1","doi-asserted-by":"crossref","unstructured":"Abadi, M., Chu, A., Goodfellow, I. J., McMahan, H. B., Mironov, I., Talwar, K., & Zhang, L. (2016). Deep Learning with Differential Privacy. In Proceedings of the 2016 ACM SIGSAC conference on computer and communications security (pp. 308\u2013318).","DOI":"10.1145\/2976749.2978318"},{"key":"10.1016\/j.ipm.2026.104879_b2","doi-asserted-by":"crossref","unstructured":"Bandodkar, G., Agarwal, S., Sughosh, A. K., Singh, S., & Choi, T. (2024). \u201cAllot?\u201d is \u201cA Lot!\u201d Towards Developing More Generalized Speech Recognition System for Accessible Communication. In Thirty-eighth AAAI conference on artificial intelligence (pp. 23327\u201323334).","DOI":"10.1609\/aaai.v38i21.30381"},{"key":"10.1016\/j.ipm.2026.104879_b3","doi-asserted-by":"crossref","unstructured":"Bu, Z., Wang, Y., Zha, S., & Karypis, G. (2023). Automatic Clipping: Differentially Private Deep Learning Made Easier and Stronger. In Advances in neural information processing systems 36: Annual conference on neural information processing systems.","DOI":"10.52202\/075280-1808"},{"issue":"7","key":"10.1016\/j.ipm.2026.104879_b4","doi-asserted-by":"crossref","first-page":"5827","DOI":"10.1109\/JIOT.2019.2952146","article-title":"Local differential privacy for deep learning","volume":"7","author":"Chamikara","year":"2020","journal-title":"IEEE Internet of Things Journal"},{"key":"10.1016\/j.ipm.2026.104879_b5","doi-asserted-by":"crossref","unstructured":"Cheng, A., Wang, J., Zhang, X. S., Chen, Q., Wang, P., & Cheng, J. (2022). DPNAS: Neural Architecture Search for Deep Learning with Differential Privacy. In Thirty-sixth AAAI conference on artificial intelligence (pp. 6358\u20136366).","DOI":"10.1609\/aaai.v36i6.20586"},{"key":"10.1016\/j.ipm.2026.104879_b6","doi-asserted-by":"crossref","unstructured":"Cummings, R., Kaptchuk, G., & Redmiles, E. M. (2021). \u201cI need a better description\u201d: An Investigation Into User Expectations For Differential Privacy. In 2021 ACM SIGSAC conference on computer and communications security (pp. 3037\u20133052).","DOI":"10.1145\/3460120.3485252"},{"key":"10.1016\/j.ipm.2026.104879_b7","first-page":"1","article-title":"Differential privacy","volume":"vol. 4052","author":"Dwork","year":"2006"},{"key":"10.1016\/j.ipm.2026.104879_b8","series-title":"Theory of cryptography, third theory of cryptography conference","first-page":"265","article-title":"Calibrating noise to sensitivity in private data analysis","author":"Dwork","year":"2006"},{"key":"10.1016\/j.ipm.2026.104879_b9","doi-asserted-by":"crossref","unstructured":"Feng, C., Xu, N., Wen, W., Venkitasubramaniam, P., & Ding, C. (2023). Spectral-DP: Differentially Private Deep Learning through Spectral Perturbation and Filtering. In 44th IEEE symposium on security and privacy (pp. 1944\u20131960).","DOI":"10.1109\/SP46215.2023.10179457"},{"issue":"6","key":"10.1016\/j.ipm.2026.104879_b10","doi-asserted-by":"crossref","first-page":"1200","DOI":"10.14778\/3648160.3648164","article-title":"DPSUR: Accelerating differentially private stochastic gradient descent using selective update and release","volume":"17","author":"Fu","year":"2024","journal-title":"Proceedings of the VLDB Endowment"},{"issue":"5","key":"10.1016\/j.ipm.2026.104879_b11","article-title":"Mitigating privacy risks in retrieval-augmented generation via locally private entity perturbation","volume":"62","author":"He","year":"2025","journal-title":"Information Processing & Management"},{"key":"10.1016\/j.ipm.2026.104879_b12","doi-asserted-by":"crossref","unstructured":"Huang, J., Wu, Q., Ren, Y., Yang, F., Yang, A., Yang, Q., & Pu, X. (2024). Sparse Bayesian Deep Learning for Cross Domain Medical Image Reconstruction. In Thirty-eighth AAAI conference on artificial intelligence (pp. 2339\u20132347).","DOI":"10.1609\/aaai.v38i3.28008"},{"key":"10.1016\/j.ipm.2026.104879_b13","doi-asserted-by":"crossref","unstructured":"Iyengar, R., Near, J. P., Song, D., Thakkar, O., Thakurta, A., & Wang, L. (2019). Towards Practical Differentially Private Convex Optimization. In 2019 IEEE symposium on security and privacy (pp. 299\u2013316).","DOI":"10.1109\/SP.2019.00001"},{"key":"10.1016\/j.ipm.2026.104879_b14","doi-asserted-by":"crossref","unstructured":"Kasiviswanathan, S. P., Lee, H. K., Nissim, K., Raskhodnikova, S., & Smith, A. D. (2008). What Can We Learn Privately?. In 49th annual IEEE symposium on foundations of computer science (pp. 531\u2013540).","DOI":"10.1109\/FOCS.2008.27"},{"key":"10.1016\/j.ipm.2026.104879_b15","series-title":"Learning multiple layers of features from tiny images","author":"Krizhevsky","year":"2009"},{"key":"10.1016\/j.ipm.2026.104879_b16","series-title":"MNIST handwritten digit database","author":"LeCun","year":"1998"},{"key":"10.1016\/j.ipm.2026.104879_b17","doi-asserted-by":"crossref","unstructured":"L\u00e9cuyer, M., Atlidakis, V., Geambasu, R., Hsu, D., & Jana, S. (2019). Certified Robustness to Adversarial Examples with Differential Privacy. In 2019 IEEE symposium on security and privacy (pp. 656\u2013672).","DOI":"10.1109\/SP.2019.00044"},{"key":"10.1016\/j.ipm.2026.104879_b18","doi-asserted-by":"crossref","unstructured":"Lee, J., & Kifer, D. (2018). Concentrated Differentially Private Gradient Descent with Adaptive per-Iteration Privacy Budget. In Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining (pp. 1656\u20131665).","DOI":"10.1145\/3219819.3220076"},{"issue":"4","key":"10.1016\/j.ipm.2026.104879_b19","doi-asserted-by":"crossref","DOI":"10.1016\/j.ipm.2024.103720","article-title":"Matrix factorization recommender based on adaptive Gaussian differential privacy for implicit feedback","volume":"61","author":"Liu","year":"2024","journal-title":"Information Processing & Management"},{"key":"10.1016\/j.ipm.2026.104879_b20","doi-asserted-by":"crossref","first-page":"2151","DOI":"10.1109\/TIFS.2022.3169911","article-title":"A differentially private framework for deep learning with convexified loss functions","volume":"17","author":"Lu","year":"2022","journal-title":"IEEE Transactions on Information Forensics Security"},{"key":"10.1016\/j.ipm.2026.104879_b21","series-title":"30th IEEE computer security foundations symposium","first-page":"263","article-title":"R\u00e9nyi differential privacy","author":"Mironov","year":"2017"},{"key":"10.1016\/j.ipm.2026.104879_b22","series-title":"2019 IEEE symposium on security and privacy","first-page":"739","article-title":"Comprehensive privacy analysis of deep learning: Passive and active white-box inference attacks against centralized and federated learning","author":"Nasr","year":"2019"},{"key":"10.1016\/j.ipm.2026.104879_b23","doi-asserted-by":"crossref","DOI":"10.1016\/j.neucom.2024.127663","article-title":"Differential privacy in deep learning: A literature survey","volume":"589","author":"Pan","year":"2024","journal-title":"Neurocomputing"},{"key":"10.1016\/j.ipm.2026.104879_b24","doi-asserted-by":"crossref","unstructured":"Papernot, N., Thakurta, A., Song, S., Chien, S., & Erlingsson, \u00da. (2021). Tempered Sigmoid Activations for Deep Learning with Differential Privacy. In Thirty-fifth AAAI conference on artificial intelligence (pp. 9312\u20139321).","DOI":"10.1609\/aaai.v35i10.17123"},{"issue":"2","key":"10.1016\/j.ipm.2026.104879_b25","doi-asserted-by":"crossref","first-page":"419","DOI":"10.1137\/22M1538703","article-title":"Differentially private sampling from distributions","volume":"54","author":"Raskhodnikova","year":"2025","journal-title":"SIAM Journal on Computing"},{"key":"10.1016\/j.ipm.2026.104879_b26","doi-asserted-by":"crossref","unstructured":"Redberg, R., Koskela, A., & Wang, Y. (2023). Improving the Privacy and Practicality of Objective Perturbation for Differentially Private Linear Learners. In Advances in neural information processing systems 36: Annual conference on neural information processing systems.","DOI":"10.52202\/075280-0609"},{"key":"10.1016\/j.ipm.2026.104879_b27","unstructured":"Salem, A., Zhang, Y., Humbert, M., Berrang, P., Fritz, M., & Backes, M. ML-Leaks: Model and Data Independent Membership Inference Attacks and Defenses on Machine Learning Models. In 26th annual network and distributed system security symposium."},{"issue":"3","key":"10.1016\/j.ipm.2026.104879_b28","doi-asserted-by":"crossref","DOI":"10.1016\/j.ipm.2023.103349","article-title":"Data release for machine learning via correlated differential privacy","volume":"60","author":"Shen","year":"2023","journal-title":"Information Processing & Management"},{"key":"10.1016\/j.ipm.2026.104879_b29","unstructured":"Tram\u00e8r, F., & Boneh, D. (2021). Differentially Private Learning Needs Better Features (or Much More Data). In 9th international conference on learning representations."},{"key":"10.1016\/j.ipm.2026.104879_b30","doi-asserted-by":"crossref","unstructured":"Wang, N., Xiao, X., Yang, Y., Zhao, J., Hui, S. C., Shin, H., Shin, J., & Yu, G. (2019). Collecting and Analyzing Multidimensional Data with Local Differential Privacy. In 35th IEEE international conference on data engineering (pp. 638\u2013649).","DOI":"10.1109\/ICDE.2019.00063"},{"issue":"1","key":"10.1016\/j.ipm.2026.104879_b31","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/TDSC.2023.3340178","article-title":"Differentially private numerical vector analyses in the local and shuffle model","volume":"22","author":"Wang","year":"2025","journal-title":"IEEE Transactions on Dependable and Secure Computing"},{"key":"10.1016\/j.ipm.2026.104879_b32","unstructured":"Wang, Z., Zhu, R., Zhou, D., Zhang, Z., Mitchell, J., Tang, H., & Wang, X. (2024). DPAdapter: Improving Differentially Private Deep Learning through Noise Tolerance Pre-training. In 33rd USeNIX security symposium."},{"key":"10.1016\/j.ipm.2026.104879_b33","doi-asserted-by":"crossref","unstructured":"Wei, J., Bao, E., Xiao, X., & Yang, Y. (2022). DPIS: An Enhanced Mechanism for Differentially Private SGD with Importance Sampling. In Proceedings of the 2022 ACM SIGSAC conference on computer and communications security (pp. 2885\u20132899).","DOI":"10.1145\/3548606.3560562"},{"key":"10.1016\/j.ipm.2026.104879_b34","doi-asserted-by":"crossref","first-page":"4498","DOI":"10.1109\/TIFS.2025.3557755","article-title":"DC-SGD: Differentially private SGD with dynamic clipping through gradient norm distribution estimation","volume":"20","author":"Wei","year":"2025","journal-title":"IEEE Transactions on Information Forensics Security"},{"key":"10.1016\/j.ipm.2026.104879_b35","series-title":"Fashion-MNIST: A novel image dataset for benchmarking machine learning algorithms","author":"Xiao","year":"2017"},{"key":"10.1016\/j.ipm.2026.104879_b36","doi-asserted-by":"crossref","unstructured":"Xiao, H., Wan, J., & Devadas, S. (2023). Geometry of Sensitivity: Twice Sampling and Hybrid Clipping in Differential Privacy with Optimal Gaussian Noise and Application to Deep Learning. In Proceedings of the 2023 ACM SIGSAC conference on computer and communications security (pp. 2636\u20132650).","DOI":"10.1145\/3576915.3623142"},{"key":"10.1016\/j.ipm.2026.104879_b37","doi-asserted-by":"crossref","first-page":"737","DOI":"10.1109\/TIFS.2024.3518069","article-title":"Privacy-preserving localization for underwater acoustic sensor networks: A differential privacy-based deep learning approach","volume":"20","author":"Yan","year":"2025","journal-title":"IEEE Transactions on Information Forensics Security"},{"key":"10.1016\/j.ipm.2026.104879_b38","doi-asserted-by":"crossref","unstructured":"Zhang, S., Hu, H., Ye, Q., & Xu, J. (2025). PrivDPR: Synthetic Graph Publishing with Deep PageRank under Differential Privacy. In Proceedings of the 31st ACM SIGKDD conference on knowledge discovery and data mining, v.1 (pp. 1936\u20131947).","DOI":"10.1145\/3690624.3709334"},{"key":"10.1016\/j.ipm.2026.104879_b39","doi-asserted-by":"crossref","unstructured":"Zhou, S., Jia, J., Wu, Z., Yang, Z., Wang, Y., Chen, W., Meng, F., Huang, S., Shen, J., & Wang, X. (2021). Inferring Emotion from Large-scale Internet Voice Data: A Semi-supervised Curriculum Augmentation based Deep Learning Approach. In Thirty-fifth AAAI conference on artificial intelligence (pp. 6039\u20136047).","DOI":"10.1609\/aaai.v35i7.16753"},{"issue":"3","key":"10.1016\/j.ipm.2026.104879_b40","first-page":"1941","article-title":"A differentially private federated learning model against poisoning attacks in edge computing","volume":"20","author":"Zhou","year":"2023","journal-title":"IEEE Transactions on Dependable and Secure Computing"},{"key":"10.1016\/j.ipm.2026.104879_b41","doi-asserted-by":"crossref","first-page":"1905","DOI":"10.1109\/TIFS.2022.3174394","article-title":"PFLF: Privacy-preserving federated learning framework for edge computing","volume":"17","author":"Zhou","year":"2022","journal-title":"IEEE Transactions on Information Forensics Security"}],"container-title":["Information Processing & Management"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0306457326002700?httpAccept=text\/xml","content-type":"text\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0306457326002700?httpAccept=text\/plain","content-type":"text\/plain","content-version":"vor","intended-application":"text-mining"}],"deposited":{"date-parts":[[2026,7,10]],"date-time":"2026-07-10T14:42:09Z","timestamp":1783694529000},"score":1,"resource":{"primary":{"URL":"https:\/\/linkinghub.elsevier.com\/retrieve\/pii\/S0306457326002700"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,11]]},"references-count":41,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2026,11]]}},"alternative-id":["S0306457326002700"],"URL":"https:\/\/doi.org\/10.1016\/j.ipm.2026.104879","relation":{},"ISSN":["0306-4573"],"issn-type":[{"value":"0306-4573","type":"print"}],"subject":[],"published":{"date-parts":[[2026,11]]},"assertion":[{"value":"Elsevier","name":"publisher","label":"This article is maintained by"},{"value":"When input perturbation outperforms gradient perturbation: Achieving high-accuracy deep learning under local differential privacy","name":"articletitle","label":"Article Title"},{"value":"Information Processing & Management","name":"journaltitle","label":"Journal Title"},{"value":"https:\/\/doi.org\/10.1016\/j.ipm.2026.104879","name":"articlelink","label":"CrossRef DOI link to publisher maintained version"},{"value":"article","name":"content_type","label":"Content Type"},{"value":"\u00a9 2026 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.","name":"copyright","label":"Copyright"}],"article-number":"104879"}}