{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,26]],"date-time":"2025-11-26T13:31:56Z","timestamp":1764163916266,"version":"3.46.0"},"reference-count":59,"publisher":"Association for Computing Machinery (ACM)","issue":"4","funder":[{"name":"National Key Research and Development Program of China","award":["2022YFB4502001"],"award-info":[{"award-number":["2022YFB4502001"]}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["62072204"],"award-info":[{"award-number":["62072204"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Auton. Adapt. Syst."],"published-print":{"date-parts":[[2025,12,31]]},"abstract":"<jats:p>\n                    Due to the limited resources of end devices, the task of\n                    <jats:italic toggle=\"yes\">Convolutional Neural Network<\/jats:italic>\n                    (CNN) inference on the end-side is moving towards edge-end collaboration. However, existing collaborative methods mainly focus on offloading CNN inference tasks from end devices to a single-edge server, which leads to inefficient use of computational resources among nearby edge servers. Moreover, offloading the CNN inference task to a single third-party server may raise privacy concerns. To address these challenges, we propose a framework named AdapCP that introduces a collaborative and adaptive parallel acceleration strategy that utilizes the end device and multiple edge servers. AdapCP consists of two stages: (1) offloading to nearby servers and (2) parallel processing of the CNN inference. For the offloading phase, we use integer linear programming to find the partition points at the inter-layer level. For the parallel phase, we first investigate intra-layer structural splitting methods tailored for both convolutional and fully connected layers. Then, we employ a\n                    <jats:italic toggle=\"yes\">Deep Deterministic Policy Gradient<\/jats:italic>\n                    (DDPG) algorithm based on the Dirichlet distribution to decide the partition points. Finally, we set a periodic update index to enhance AdapCP\u2019s adaptability to dynamic environments. Empirical evaluations conducted on the Jetson nano demonstrate that AdapCP significantly reduces the total latency of CNN inference by an average factor of 2.21\n                    <jats:inline-formula content-type=\"math\/tex\">\n                      <jats:tex-math notation=\"LaTeX\" version=\"MathJax\">\\(\\times\\)<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    compared to existing solutions.\n                  <\/jats:p>","DOI":"10.1145\/3765961","type":"journal-article","created":{"date-parts":[[2025,9,5]],"date-time":"2025-09-05T15:10:44Z","timestamp":1757085044000},"page":"1-28","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["AdapCP: Collaborative Inference with Adaptive CNN Partition on Distributed Edge Servers"],"prefix":"10.1145","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-7735-3187","authenticated-orcid":false,"given":"Sifan","family":"Zhao","sequence":"first","affiliation":[{"name":"Huazhong University of Science and Technology, Wuhan, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0336-0522","authenticated-orcid":false,"given":"Dezhong","family":"Yao","sequence":"additional","affiliation":[{"name":"Huazhong University of Science and Technology, Wuhan, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6937-4180","authenticated-orcid":false,"given":"Yao","family":"Wan","sequence":"additional","affiliation":[{"name":"Huazhong University of Science and Technology, Wuhan, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6615-0699","authenticated-orcid":false,"given":"Gang","family":"Wu","sequence":"additional","affiliation":[{"name":"National Super Computing Center in Zhengzhou, Zhengzhou University, Zhengzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3934-7605","authenticated-orcid":false,"given":"Hai","family":"Jin","sequence":"additional","affiliation":[{"name":"Huazhong University of Science and Technology, Wuhan, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2025,11,26]]},"reference":[{"key":"e_1_3_2_2_2","article-title":"Faster R-CNN: Towards real-time object detection with region proposal networks","volume":"28","author":"Ren Shaoqing","year":"2015","unstructured":"Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. 2015. Faster R-CNN: Towards real-time object detection with region proposal networks. In Proceedings of the Advances in Neural Information Processing Systems, Vol. 28.","journal-title":"Proceedings of the Advances in Neural Information Processing Systems"},{"key":"e_1_3_2_3_2","doi-asserted-by":"publisher","DOI":"10.1109\/WACV57701.2024.00682"},{"key":"e_1_3_2_4_2","doi-asserted-by":"publisher","DOI":"10.1109\/TASLP.2020.3036237"},{"key":"e_1_3_2_5_2","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2024.3366506"},{"key":"e_1_3_2_6_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCE.2024.3436908"},{"key":"e_1_3_2_7_2","doi-asserted-by":"publisher","DOI":"10.1145\/3495243.3560551"},{"key":"e_1_3_2_8_2","doi-asserted-by":"publisher","DOI":"10.1145\/3582080"},{"key":"e_1_3_2_9_2","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM41043.2020.9155284"},{"issue":"7","key":"e_1_3_2_10_2","first-page":"1539","article-title":"Accelerating federated learning over reliability-agnostic clients in mobile edge computing systems","volume":"32","author":"Wu Wentai","year":"2020","unstructured":"Wentai Wu, Ligang He, Weiwei Lin, and Rui Mao. 2020. Accelerating federated learning over reliability-agnostic clients in mobile edge computing systems. IEEE Transactions on Parallel and Distributed Systems 32, 7 (2020), 1539\u20131551.","journal-title":"IEEE Transactions on Parallel and Distributed Systems"},{"key":"e_1_3_2_11_2","doi-asserted-by":"publisher","DOI":"10.1109\/TSUSC.2022.3216461"},{"key":"e_1_3_2_12_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-022-3652-2"},{"key":"e_1_3_2_13_2","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2019.2947893"},{"key":"e_1_3_2_14_2","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2019.2893250"},{"key":"e_1_3_2_15_2","doi-asserted-by":"publisher","DOI":"10.1145\/3450268.3453520"},{"key":"e_1_3_2_16_2","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOMWKSHPS47286.2019.9093772"},{"key":"e_1_3_2_17_2","doi-asserted-by":"publisher","DOI":"10.3390\/app12189010"},{"key":"e_1_3_2_18_2","doi-asserted-by":"publisher","DOI":"10.1186\/s13677-023-00404-y"},{"key":"e_1_3_2_19_2","doi-asserted-by":"publisher","DOI":"10.1145\/3595292"},{"key":"e_1_3_2_20_2","doi-asserted-by":"publisher","DOI":"10.1145\/3590154"},{"key":"e_1_3_2_21_2","doi-asserted-by":"publisher","DOI":"10.1145\/3487920"},{"key":"e_1_3_2_22_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVT.2023.3309321"},{"key":"e_1_3_2_23_2","doi-asserted-by":"publisher","DOI":"10.1016\/S0096-3003(97)10140-0"},{"key":"e_1_3_2_24_2","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2025.3562721"},{"key":"e_1_3_2_25_2","doi-asserted-by":"crossref","unstructured":"Thanh-Nghi Do and Minh-Thu Tran-Nguyen. 2024. ImageNet classification with Raspberry Pis: Federated learning algorithms of local classifiers. International Journal of Web Information Systems 20 (2024) 48\u201365.","DOI":"10.1108\/IJWIS-03-2023-0057"},{"key":"e_1_3_2_26_2","doi-asserted-by":"publisher","DOI":"10.1145\/3093337.3037698"},{"key":"e_1_3_2_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2021.3100298"},{"key":"e_1_3_2_28_2","doi-asserted-by":"publisher","DOI":"10.1109\/RTAS48715.2020.00-20"},{"key":"e_1_3_2_29_2","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2021.3075464"},{"key":"e_1_3_2_30_2","doi-asserted-by":"publisher","DOI":"10.1145\/3267809.3267828"},{"key":"e_1_3_2_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS47774.2020.00114"},{"key":"e_1_3_2_32_2","doi-asserted-by":"publisher","DOI":"10.1145\/3543507.3583306"},{"key":"e_1_3_2_33_2","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2023.3290145"},{"key":"e_1_3_2_34_2","doi-asserted-by":"publisher","DOI":"10.1145\/3359789.3359824"},{"key":"e_1_3_2_35_2","doi-asserted-by":"publisher","DOI":"10.1109\/TNSE.2022.3177755"},{"key":"e_1_3_2_36_2","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2020.3022358"},{"key":"e_1_3_2_37_2","doi-asserted-by":"crossref","unstructured":"Yulong Wang Xingshu Chen and Qixu Wang. 2022. Privacy-preserving security inference towards cloud-edge collaborative using differential privacy. arXiv:2212.06428. Retrieved from https:\/\/arxiv.org\/abs\/2212.06428","DOI":"10.2139\/ssrn.4581848"},{"key":"e_1_3_2_38_2","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2023.3312304"},{"key":"e_1_3_2_39_2","unstructured":"Shiwei Ding Lan Zhang Miao Pan and Xiaoyong Yuan. 2023. PATROL: Privacy-oriented pruning for collaborative inference against model inversion attacks. arXiv:2307.10981. Retrieved from https:\/\/arxiv.org\/abs\/2307.10981"},{"key":"e_1_3_2_40_2","doi-asserted-by":"crossref","unstructured":"Wenjie Hou and Ying Chen. 2025. User allocation in multi-hop edge computing networks: A game-theoretical approach. International Journal of Web Information Systems 21 (2025) 230\u2013253.","DOI":"10.1108\/IJWIS-07-2024-0190"},{"key":"e_1_3_2_41_2","first-page":"1","article-title":"Reinforcement learning based edge-end collaboration for multi-task scheduling in 6G enabled intelligent autonomous transport systems","author":"Li Peisong","year":"2025","unstructured":"Peisong Li, Ziren Xiao, Honghao Gao, Xinheng Wang, and Ye Wang. 2025. Reinforcement learning based edge-end collaboration for multi-task scheduling in 6G enabled intelligent autonomous transport systems. IEEE Transactions on Intelligent Transportation Systems (2025), 1\u201314.","journal-title":"IEEE Transactions on Intelligent Transportation Systems"},{"key":"e_1_3_2_42_2","doi-asserted-by":"publisher","DOI":"10.1109\/TVT.2021.3068255"},{"key":"e_1_3_2_43_2","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS53621.2022.00110"},{"key":"e_1_3_2_44_2","unstructured":"Timothy P. Lillicrap Jonathan J. Hunt Alexander Pritzel Nicolas Heess Tom Erez Yuval Tassa David Silver and Daan Wierstra. 2015. Continuous control with deep reinforcement learning. arXiv:1509.02971. Retrieved from https:\/\/arxiv.org\/abs\/1509.02971"},{"key":"e_1_3_2_45_2","doi-asserted-by":"publisher","DOI":"10.5555\/1074100.1074233"},{"key":"e_1_3_2_46_2","doi-asserted-by":"publisher","DOI":"10.1145\/3703352"},{"key":"e_1_3_2_47_2","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM41043.2020.9155237"},{"key":"e_1_3_2_48_2","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2020.3042320"},{"key":"e_1_3_2_49_2","doi-asserted-by":"publisher","DOI":"10.1038\/nature14236"},{"key":"e_1_3_2_50_2","doi-asserted-by":"publisher","DOI":"10.1007\/BF00992699"},{"key":"e_1_3_2_51_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10462-021-10061-9"},{"key":"e_1_3_2_52_2","doi-asserted-by":"publisher","DOI":"10.1109\/GCWkshps45667.2019.9024603"},{"key":"e_1_3_2_53_2","doi-asserted-by":"publisher","DOI":"10.1023\/A:1022633531479"},{"key":"e_1_3_2_54_2","doi-asserted-by":"publisher","DOI":"10.1109\/TIFS.2024.3384841"},{"key":"e_1_3_2_55_2","doi-asserted-by":"publisher","DOI":"10.1145\/3595292"},{"key":"e_1_3_2_56_2","unstructured":"Karen Simonyan and Andrew Zisserman. 2014. Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556. Retrieved from https:\/\/arxiv.org\/abs\/1409.1556"},{"key":"e_1_3_2_57_2","unstructured":"Andrew G. Howard Menglong Zhu Bo Chen Dmitry Kalenichenko Weijun Wang Tobias Weyand Marco Andreetto and Hartwig Adam. 2017. MobileNets: Efficient convolutional neural networks for mobile vision applications. arXiv:1704.04861. Retrieved from https:\/\/arxiv.org\/abs\/1704.04861"},{"key":"e_1_3_2_58_2","doi-asserted-by":"publisher","DOI":"10.1145\/3065386"},{"key":"e_1_3_2_59_2","unstructured":"Min Lin Qiang Chen and Shuicheng Yan. 2013. Network in network. arXiv:1312.4400. Retrieved from https:\/\/arxiv.org\/abs\/1312.4400"},{"key":"e_1_3_2_60_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2022.10.033"}],"container-title":["ACM Transactions on Autonomous and Adaptive Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3765961","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,26]],"date-time":"2025-11-26T13:29:04Z","timestamp":1764163744000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3765961"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,26]]},"references-count":59,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2025,12,31]]}},"alternative-id":["10.1145\/3765961"],"URL":"https:\/\/doi.org\/10.1145\/3765961","relation":{},"ISSN":["1556-4665","1556-4703"],"issn-type":[{"type":"print","value":"1556-4665"},{"type":"electronic","value":"1556-4703"}],"subject":[],"published":{"date-parts":[[2025,11,26]]},"assertion":[{"value":"2024-02-20","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-08-23","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-11-26","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}