{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,29]],"date-time":"2026-05-29T17:00:43Z","timestamp":1780074043092,"version":"3.54.0"},"reference-count":36,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/tdm\/userlicense\/1.0\/"},{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/legal\/tdmrep-license"},{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-017"},{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-012"},{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,12,1]],"date-time":"2025-12-01T00:00:00Z","timestamp":1764547200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-004"}],"funder":[{"DOI":"10.13039\/501100003392","name":"Fujian Provincial Natural Science Foundation","doi-asserted-by":"publisher","award":["2025J01537"],"award-info":[{"award-number":["2025J01537"]}],"id":[{"id":"10.13039\/501100003392","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["U21A20471"],"award-info":[{"award-number":["U21A20471"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["elsevier.com","sciencedirect.com"],"crossmark-restriction":true},"short-container-title":["Journal of Systems Architecture"],"published-print":{"date-parts":[[2025,12]]},"DOI":"10.1016\/j.sysarc.2025.103593","type":"journal-article","created":{"date-parts":[[2025,10,9]],"date-time":"2025-10-09T07:20:40Z","timestamp":1759994440000},"page":"103593","update-policy":"https:\/\/doi.org\/10.1016\/elsevier_cm_policy","source":"Crossref","is-referenced-by-count":2,"special_numbering":"C","title":["Fault resilient on-device batched DNN inference for mobile devices with ARM TrustZone"],"prefix":"10.1016","volume":"169","author":[{"given":"Longlong","family":"Liao","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wenbin","family":"Zeng","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xinqi","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jie","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuanlong","family":"Yu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"key":"10.1016\/j.sysarc.2025.103593_b1","doi-asserted-by":"crossref","unstructured":"A.M. Azab, P. Ning, J. Shah, Q. Chen, R. Bhutkar, G. Ganesh, J. Ma, W. Shen, Hypervision across worlds: Real-time kernel protection from the ARM TrustZone secure world, in: Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security, ACM CCS, 2014, pp. 90\u2013102.","DOI":"10.1145\/2660267.2660350"},{"key":"10.1016\/j.sysarc.2025.103593_b2","series-title":"Confidential computing on NVIDIA Hopper GPUs: A performance benchmark study","author":"Zhu","year":"2024"},{"issue":"9","key":"10.1016\/j.sysarc.2025.103593_b3","doi-asserted-by":"crossref","first-page":"2952","DOI":"10.1109\/TCAD.2023.3235803","article-title":"AOS: An automated overclocking system for high-performance CNN accelerator through timing delay measurement on FPGA","volume":"42","author":"Jiang","year":"2023","journal-title":"IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst."},{"issue":"6","key":"10.1016\/j.sysarc.2025.103593_b4","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3291047","article-title":"Demystifying ARM TrustZone: A comprehensive survey","volume":"51","author":"Pinto","year":"2019","journal-title":"ACM Comput. Surv."},{"key":"10.1016\/j.sysarc.2025.103593_b5","series-title":"2021 IEEE\/ACM Symposium on Edge Computing","first-page":"441","article-title":"TrustZone enhanced plausibly deniable encryption system for mobile devices","author":"Liao","year":"2021"},{"key":"10.1016\/j.sysarc.2025.103593_b6","series-title":"2020 IEEE Global Communications Conference","first-page":"1","article-title":"On runtime software security of trustzone-m based IoT devices","author":"Luo","year":"2020"},{"issue":"15","key":"10.1016\/j.sysarc.2025.103593_b7","doi-asserted-by":"crossref","DOI":"10.1002\/cpe.6936","article-title":"Darknet on OpenCL: A multiplatform tool for object detection and classification","volume":"34","author":"Sowa","year":"2022","journal-title":"Concurr. Comput.: Pr. Exp."},{"key":"10.1016\/j.sysarc.2025.103593_b8","series-title":"2024 IEEE International Symposium on Hardware Oriented Security and Trust","first-page":"01","article-title":"TrustZoneTunnel: A cross-world pattern history table-based microarchitectural side-channel attack","author":"Xu","year":"2024"},{"key":"10.1016\/j.sysarc.2025.103593_b9","series-title":"IEEE Conference on Computer Communications","first-page":"2009","article-title":"Memory-efficient and secure DNN inference on TrustZone-enabled consumer IoT devices","author":"Xie","year":"2024"},{"key":"10.1016\/j.sysarc.2025.103593_b10","series-title":"2022 IEEE Symposium on Security and Privacy","first-page":"352","article-title":"Rt-tee: Real-time system availability for cyber-physical systems using ARM TrustZone","author":"Wang","year":"2022"},{"key":"10.1016\/j.sysarc.2025.103593_b11","doi-asserted-by":"crossref","unstructured":"F. Mo, A.S. Shamsabadi, K. Katevas, et al., DarkneTZ: Towards model privacy at the edge using Trusted Execution Environments, in: Proceedings of the 18th International Conference on Mobile Systems, Applications, and Services, MobiSys, 2020, pp. 161\u2013174.","DOI":"10.1145\/3386901.3388946"},{"key":"10.1016\/j.sysarc.2025.103593_b12","series-title":"2023 32nd International Conference on Parallel Architectures and Compilation Techniques","first-page":"187","article-title":"Automatic algorithm-based fault tolerance (AABFT) of stencil computations","author":"Narmour","year":"2023"},{"issue":"4","key":"10.1016\/j.sysarc.2025.103593_b13","doi-asserted-by":"crossref","first-page":"2546","DOI":"10.1109\/TDSC.2021.3063083","article-title":"Making convolutions resilient via algorithm-based error detection techniques","volume":"19","author":"Hari","year":"2021","journal-title":"IEEE Trans. Dependable Secur. Comput."},{"issue":"7","key":"10.1016\/j.sysarc.2025.103593_b14","first-page":"1677","article-title":"FT-CNN: Algorithm-based fault tolerance for convolutional neural networks","volume":"32","author":"Zhao","year":"2020","journal-title":"IEEE Trans. Parallel Distrib. Syst."},{"key":"10.1016\/j.sysarc.2025.103593_b15","doi-asserted-by":"crossref","unstructured":"J. Kosaian, K. Rashmi, Arithmetic-intensity-guided fault tolerance for neural network inference on GPUs, in: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, SC21, 2021, pp. 1\u201315.","DOI":"10.1145\/3458817.3476184"},{"key":"10.1016\/j.sysarc.2025.103593_b16","series-title":"Proceedings of the 61st ACM\/IEEE Design Automation Conference","first-page":"1","article-title":"Maintaining sanity: Algorithm-based comprehensive fault tolerance for CNNs","author":"Jung","year":"2024"},{"key":"10.1016\/j.sysarc.2025.103593_b17","series-title":"2021 IEEE International Symposium on High-Performance Computer Architecture","first-page":"493","article-title":"Lazy batching: An sla-aware batching system for cloud machine learning inference","author":"Choi","year":"2021"},{"key":"10.1016\/j.sysarc.2025.103593_b18","series-title":"2023 IEEE International Parallel and Distributed Processing Symposium Workshops","first-page":"326","article-title":"A high-dimensional algorithm-based fault tolerance scheme","author":"Fu","year":"2023"},{"key":"10.1016\/j.sysarc.2025.103593_b19","series-title":"2024 IEEE\/ACM 32nd International Symposium on Quality of Service","first-page":"1","article-title":"HarmonyBatch: Batching multi-SLO DNN inference with heterogeneous serverless functions","author":"Chen","year":"2024"},{"key":"10.1016\/j.sysarc.2025.103593_b20","series-title":"2024 IEEE 99th Vehicular Technology Conference","first-page":"1","article-title":"Dynamic batching and early-exiting for accurate and timely edge inference","author":"She","year":"2024"},{"issue":"10","key":"10.1016\/j.sysarc.2025.103593_b21","doi-asserted-by":"crossref","first-page":"2496","DOI":"10.1109\/TPDS.2022.3144614","article-title":"Coordinated batching and DVFS for DNN inference on GPU accelerators","volume":"33","author":"Nabavinejad","year":"2022","journal-title":"IEEE Trans. Parallel Distrib. Syst."},{"key":"10.1016\/j.sysarc.2025.103593_b22","doi-asserted-by":"crossref","unstructured":"S. Baek, M. Wi, S. Park, H. Nam, M.J. Kim, N.S. Kim, J.H. Ahn, Marionette: A rowhammer attack via row coupling, in: Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, ASPLOS 2025, Vol. 1, 2025, pp. 637\u2013652.","DOI":"10.1145\/3669940.3707242"},{"key":"10.1016\/j.sysarc.2025.103593_b23","series-title":"32nd USENIX Security Symposium","first-page":"6239","article-title":"Oops..! I glitched it again! How to multi-glitch the glitching-protections on ARM TrustZone-M","author":"Sa\u00df","year":"2023"},{"key":"10.1016\/j.sysarc.2025.103593_b24","doi-asserted-by":"crossref","unstructured":"P. Zhao, S. Wang, C. Gongye, et al., Fault sneaking attack: A stealthy framework for misleading deep neural networks, in: 56th ACM\/IEEE Design Automation Conference, DAC, 2019, pp. 1\u20136.","DOI":"10.1145\/3316781.3317825"},{"key":"10.1016\/j.sysarc.2025.103593_b25","doi-asserted-by":"crossref","unstructured":"M.R.H. Rashed, S. Thijssen, D. Simon, S. Jha, R. Ewetz, Execution Sequence Optimization for Processing In-Memory using Parallel Data Preparation, in: Proceedings of the 61st ACM\/IEEE Design Automation Conference, DAC, 2024, pp. 1\u20136.","DOI":"10.1145\/3649329.3657348"},{"key":"10.1016\/j.sysarc.2025.103593_b26","doi-asserted-by":"crossref","unstructured":"Z. Liu, T. Zhou, Y. Luo, X. Xu, TBNet: A Neural Architectural Defense Framework Facilitating DNN Model Protection in Trusted Execution Environments, in: Proceedings of the 61st ACM\/IEEE Design Automation Conference, DAC, 2024, pp. 1\u20136.","DOI":"10.1145\/3649329.3658251"},{"key":"10.1016\/j.sysarc.2025.103593_b27","doi-asserted-by":"crossref","unstructured":"H. Wang, C. Ma, An optimization of im2col, an important method of CNNs, based on continuous address access, in: IEEE International Conference on Consumer Electronics and Computer Engineering, ICCECE, 2021, pp. 314\u2013320.","DOI":"10.1109\/ICCECE51280.2021.9342343"},{"issue":"1","key":"10.1016\/j.sysarc.2025.103593_b28","doi-asserted-by":"crossref","first-page":"40","DOI":"10.1145\/363707.363723","article-title":"Pracniques: further remarks on reducing truncation errors","volume":"8","author":"Kahan","year":"1965","journal-title":"Commun. ACM"},{"key":"10.1016\/j.sysarc.2025.103593_b29","doi-asserted-by":"crossref","first-page":"279","DOI":"10.1007\/s00607-005-0139-x","article-title":"A generalized kahan-babu\u0161ka-summation-algorithm","volume":"76","author":"Klein","year":"2006","journal-title":"Computing"},{"key":"10.1016\/j.sysarc.2025.103593_b30","doi-asserted-by":"crossref","unstructured":"J. Redmon, A. Farhadi, YOLO9000: Better, faster, stronger, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, CVPR, 2017, pp. 7263\u20137271.","DOI":"10.1109\/CVPR.2017.690"},{"key":"10.1016\/j.sysarc.2025.103593_b31","doi-asserted-by":"crossref","unstructured":"K. He, X. Zhang, S. Ren, et al., Deep residual learning for image recognition, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, CVPR, 2016, pp. 770\u2013778.","DOI":"10.1109\/CVPR.2016.90"},{"key":"10.1016\/j.sysarc.2025.103593_b32","doi-asserted-by":"crossref","unstructured":"G. Huang, Z. Liu, L. Van Der Maaten, K.Q. Weinberger, Densely connected convolutional networks, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, CVPR, 2017, pp. 4700\u20134708.","DOI":"10.1109\/CVPR.2017.243"},{"key":"10.1016\/j.sysarc.2025.103593_b33","series-title":"YOLOv3: An incremental improvement","author":"Redmon","year":"2018"},{"key":"10.1016\/j.sysarc.2025.103593_b34","doi-asserted-by":"crossref","unstructured":"C.Y. Wang, A. Bochkovskiy, H.Y.M. Liao, Scaled-YOLOv4: Scaling Cross Stage Partial Network, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, CVPR, 2021, pp. 13024\u201313033.","DOI":"10.1109\/CVPR46437.2021.01283"},{"key":"10.1016\/j.sysarc.2025.103593_b35","series-title":"Microsoft COCO captions: Data collection and evaluation server","author":"Chen","year":"2015"},{"key":"10.1016\/j.sysarc.2025.103593_b36","doi-asserted-by":"crossref","first-page":"211","DOI":"10.1007\/s11263-015-0816-y","article-title":"ImageNet large scale visual recognition challenge","volume":"115","author":"Russakovsky","year":"2015","journal-title":"Int. J. Comput. Vis."}],"container-title":["Journal of Systems Architecture"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S1383762125002656?httpAccept=text\/xml","content-type":"text\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S1383762125002656?httpAccept=text\/plain","content-type":"text\/plain","content-version":"vor","intended-application":"text-mining"}],"deposited":{"date-parts":[[2026,3,10]],"date-time":"2026-03-10T15:26:27Z","timestamp":1773156387000},"score":1,"resource":{"primary":{"URL":"https:\/\/linkinghub.elsevier.com\/retrieve\/pii\/S1383762125002656"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12]]},"references-count":36,"alternative-id":["S1383762125002656"],"URL":"https:\/\/doi.org\/10.1016\/j.sysarc.2025.103593","relation":{},"ISSN":["1383-7621"],"issn-type":[{"value":"1383-7621","type":"print"}],"subject":[],"published":{"date-parts":[[2025,12]]},"assertion":[{"value":"Elsevier","name":"publisher","label":"This article is maintained by"},{"value":"Fault resilient on-device batched DNN inference for mobile devices with ARM TrustZone","name":"articletitle","label":"Article Title"},{"value":"Journal of Systems Architecture","name":"journaltitle","label":"Journal Title"},{"value":"https:\/\/doi.org\/10.1016\/j.sysarc.2025.103593","name":"articlelink","label":"CrossRef DOI link to publisher maintained version"},{"value":"article","name":"content_type","label":"Content Type"},{"value":"\u00a9 2025 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.","name":"copyright","label":"Copyright"}],"article-number":"103593"}}