{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,3]],"date-time":"2026-07-03T09:02:43Z","timestamp":1783069363499,"version":"3.54.6"},"reference-count":47,"publisher":"Association for Computing Machinery (ACM)","issue":"3","license":[{"start":{"date-parts":[[2024,5,11]],"date-time":"2024-05-11T00:00:00Z","timestamp":1715385600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"name":"PNRR-PE-AI FAIR"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Embed. Comput. Syst."],"published-print":{"date-parts":[[2024,5,31]]},"abstract":"<jats:p>Incremental on-device learning is one of the most relevant and interesting challenges in the field of Tiny Machine Learning (TinyML). Indeed, differently from traditional TinyML solutions where the training is typically carried out on the Cloud and inference only occurs on the tiny devices (e.g., embedded systems or Internet-of-Things units), incremental on-device TinyML allows both the inference and the training of TinyML models directly on tiny devices.<\/jats:p>\n          <jats:p>This ability paves the way for TinyML-enabled intelligent devices that can learn directly on the field and adapt to evolving environments, different working conditions, or specific users. The literature in this field is quite limited with very few solutions focusing only on the incremental fine-tuning of machine learning models, whereas a general solution encompassing algorithms and code generation for incremental on-device TinyML is still perceived as missing.<\/jats:p>\n          <jats:p>\n            The aim of this article is to introduce, to the best of our knowledge for the first time in the literature, a toolbox called\n            <jats:italic>TyBox<\/jats:italic>\n            for the automatic design and code generation of incremental on-device TinyML classification models. In more detail, starting from a \u201cstatic\u201d TinyML model, TyBox is able to (i) automatically design the \u201cincremental\u201d on-device version of the TinyML model that has been suitably designed to take into account the technological constraint on the RAM memory of the target tiny device, and (ii) autonomously provide the C++ codes and libraries to support the inference and learning of the incremental on-device TinyML model directly on the tiny devices.\n          <\/jats:p>\n          <jats:p>TyBox has been extensively compared with a state-of-the-art incremental learning solution for TinyML and tested on an off-the-shelf tiny device (i.e., the Arduino Nano 33 BLE) in three relevant TinyML application tasks and scenarios: binary image classification, multi-class image classification, and ultra-wide-band human activity recognition. In addition, TyBox is released to the scientific community as a public repository.<\/jats:p>","DOI":"10.1145\/3604566","type":"journal-article","created":{"date-parts":[[2023,6,17]],"date-time":"2023-06-17T09:12:46Z","timestamp":1686993166000},"page":"1-27","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":18,"title":["TyBox: An Automatic Design and Code Generation Toolbox for TinyML Incremental On-Device Learning"],"prefix":"10.1145","volume":"23","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-5964-5685","authenticated-orcid":false,"given":"Massimo","family":"Pavan","sequence":"first","affiliation":[{"name":"Politecnico di Milano, Milano, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-4419-980X","authenticated-orcid":false,"given":"Eugeniu","family":"Ostrovan","sequence":"additional","affiliation":[{"name":"Politecnico di Milano, Milano, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-3568-0415","authenticated-orcid":false,"given":"Armando","family":"Caltabiano","sequence":"additional","affiliation":[{"name":"Truesense s.r.l., Milano, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7828-7687","authenticated-orcid":false,"given":"Manuel","family":"Roveri","sequence":"additional","affiliation":[{"name":"Politecnico di Milano, Milano, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2024,5,11]]},"reference":[{"key":"e_1_3_4_2_2","unstructured":"Arduino. n.d. Arduino Nano 33 BLE Sense. Retrieved June 21 2023 from https:\/\/docs.arduino.cc\/hardware\/nano-33-ble-sense."},{"key":"e_1_3_4_3_2","volume-title":"Banknote Authentication Dataset","author":"UC Irvine","unstructured":"UC Irvine. n.d. Banknote Authentication Dataset. https:\/\/archive.ics.uci.edu\/ml\/datasets\/banknote+authentication."},{"key":"e_1_3_4_4_2","volume-title":"Fashion MNIST Dataset","author":"GitHub","unstructured":"GitHub. n.d. Fashion MNIST Dataset. Retrieved June 21, 2023 from https:\/\/github.com\/zalandoresearch\/fashion-mnist."},{"key":"e_1_3_4_5_2","volume-title":"MNIST Dataset","author":"GitHub","unstructured":"GitHub. n.d. MNIST Dataset. Retrieved June 21, 2023 from https:\/\/storage.googleapis.com\/tensorflow\/tf-keras-datasets\/mnist.npz."},{"key":"e_1_3_4_6_2","unstructured":"Mart\u00edn Abadi Ashish Agarwal Paul Barham Eugene Brevdo Zhifeng Chen Craig Citro Greg S. Corrado et\u00a0al. 2015. TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems. Retrieved June 21 2023 from https:\/\/www.tensorflow.org."},{"key":"e_1_3_4_7_2","doi-asserted-by":"publisher","DOI":"10.1109\/IJCNN.2011.6033426"},{"key":"e_1_3_4_8_2","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2013.2239309"},{"key":"e_1_3_4_9_2","doi-asserted-by":"publisher","DOI":"10.1109\/IPSN.2018.00049"},{"key":"e_1_3_4_10_2","doi-asserted-by":"publisher","DOI":"10.1109\/MC.2017.111"},{"key":"e_1_3_4_11_2","doi-asserted-by":"publisher","DOI":"10.1109\/TBCAS.2019.2914476"},{"key":"e_1_3_4_12_2","article-title":"Tiny transfer learning: Towards memory-efficient on-device learning","author":"Cai Han","unstructured":"Han Cai, Chuang Gan, Ligeng Zhu, and Song Han. n.d. Tiny transfer learning: Towards memory-efficient on-device learning. arXiv:2007.11622 [cs] (n.d.).","journal-title":"arXiv:2007.11622 [cs]"},{"key":"e_1_3_4_13_2","unstructured":"Robert David Jared Duke Advait Jain Vijay Janapa Reddi Nat Jeffries Jian Li Nick Kreeger et\u00a0al. 2020. TensorFlow lite micro: Embedded machine learning on TinyML systems. arXiv:2010.08678 (2020)."},{"issue":"7","key":"e_1_3_4_14_2","first-page":"3366","article-title":"A continual learning survey: Defying forgetting in classification tasks","volume":"44","author":"Lange Matthias De","year":"2021","unstructured":"Matthias De Lange, Rahaf Aljundi, Marc Masana, Sarah Parisot, Xu Jia, Ale\u0161 Leonardis, Gregory Slabaugh, and Tinne Tuytelaars. 2021. A continual learning survey: Defying forgetting in classification tasks. IEEE Transactions on Pattern Analysis and Machine Intelligence 44, 7 (2021), 3366\u20133385.","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"e_1_3_4_15_2","first-page":"1","volume-title":"Proceedings of the 2018 International Joint Conference on Neural Networks (IJCNN\u201918)","author":"Disabato Simone","unstructured":"Simone Disabato and Manuel Roveri. 2018. Reducing the computation load of convolutional neural networks through gate classification. In Proceedings of the 2018 International Joint Conference on Neural Networks (IJCNN\u201918). IEEE, Los Alamitos, CA, 1\u20138."},{"key":"e_1_3_4_16_2","doi-asserted-by":"crossref","unstructured":"Simone Disabato and Manuel Roveri. 2020. Incremental on-device tiny machine learning. In Proceedings ofthe 2nd International Workshop on Challenges in Artificial Intelligence and Machine Learning for Internet of Things(AIChallengIoT\u201920). 7\u201313.","DOI":"10.1145\/3417313.3429378"},{"key":"e_1_3_4_17_2","article-title":"Tiny machine learning for concept drift","author":"Disabato Simone","year":"2021","unstructured":"Simone Disabato and Manuel Roveri. 2021. Tiny machine learning for concept drift. arXiv:2107.14759 [cs] (2021). http:\/\/arxiv.org\/abs\/2107.14759.","journal-title":"arXiv:2107.14759 [cs]"},{"key":"e_1_3_4_18_2","doi-asserted-by":"publisher","DOI":"10.1109\/MCI.2015.2471196"},{"key":"e_1_3_4_19_2","doi-asserted-by":"publisher","DOI":"10.1109\/MCI.2015.2471196"},{"key":"e_1_3_4_20_2","doi-asserted-by":"crossref","unstructured":"Joao Gama Pedro Medas Gladys Castillo and Pedro Rodrigues. 2004. Learning with drift detection. In Advances in Artificial Intelligence\u2014SBIA 2004. Lecture Notes in Computer Science Vol. 3171. Springer 286\u2013295.","DOI":"10.1007\/978-3-540-28645-5_29"},{"key":"e_1_3_4_21_2","volume-title":"Proceedings of the European Symposium on Artificial Neural Networks (ESANN\u201916)","author":"Gepperth Alexander","year":"2016","unstructured":"Alexander Gepperth and Barbara Hammer. 2016. Incremental learning algorithms and applications. In Proceedings of the European Symposium on Artificial Neural Networks (ESANN\u201916)."},{"key":"e_1_3_4_22_2","doi-asserted-by":"crossref","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:2103.13630 [cs] (2021).","DOI":"10.1201\/9781003162810-13"},{"key":"e_1_3_4_23_2","first-page":"249","volume-title":"Proceedings of the 13th International Conference on Artificial Intelligence and Statistics","author":"Glorot Xavier","year":"2010","unstructured":"Xavier Glorot and Yoshua Bengio. 2010. Understanding the difficulty of training deep feedforward neural networks. In Proceedings of the 13th International Conference on Artificial Intelligence and Statistics. 249\u2013256."},{"key":"e_1_3_4_24_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10710-017-9314-z"},{"key":"e_1_3_4_25_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 (2017)."},{"key":"e_1_3_4_26_2","unstructured":"Forrest N. Iandola Song Han Matthew W. Moskewicz Khalid Ashraf William J. Dally and Kurt Keutzer. 2016. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size. arXiv:1602.07360 (2016)."},{"key":"e_1_3_4_27_2","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1611835114"},{"key":"e_1_3_4_28_2","unstructured":"Ji Lin Ligeng Zhu Wei-Ming Chen Wei-Chen Wang Chuang Gan and Song Han. 2022. On-d training under 256KB memory. arXiv:2206.15472 [cs] (2022). http:\/\/arxiv.org\/abs\/2206.15472."},{"key":"e_1_3_4_29_2","unstructured":"Jiayi Liu Samarth Tripathy Unmesh Kurup and Mohak Shah. 2020. Pruning algorithms to accelerate convolutional neural networks for edge applications: A survey. arXiv:2005.04275 [cs stat] (2020)."},{"issue":"12","key":"e_1_3_4_30_2","first-page":"2346","article-title":"Learning under concept drift: A review","volume":"31","author":"Lu Jie","year":"2018","unstructured":"Jie Lu, Anjin Liu, Fan Dong, Feng Gu, Joao Gama, and Guangquan Zhang. 2018. Learning under concept drift: A review. IEEE Transactions on Knowledge and Data Engineering 31, 12 (2018), 2346\u20132363.","journal-title":"IEEE Transactions on Knowledge and Data Engineering"},{"key":"e_1_3_4_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00810"},{"key":"e_1_3_4_32_2","unstructured":"Marc Masana Xialei Liu Bartlomiej Twardowski Mikel Menta Andrew D. Bagdanov and Joost van de Weijer. 2020. Class-incremental learning: Survey and performance evaluation on image classification. arXiv preprint arXiv:2010.15277 (2020)."},{"key":"e_1_3_4_33_2","doi-asserted-by":"crossref","unstructured":"Michael McCloskey and Neal J. Cohen. 1989. Catastrophic interference in connectionist networks: The sequential learning problem. Psychology of Learning and Motivation24 (1989) 109\u2013165.","DOI":"10.1016\/S0079-7421(08)60536-8"},{"key":"e_1_3_4_34_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2019.01.012"},{"key":"e_1_3_4_35_2","doi-asserted-by":"crossref","unstructured":"Lorenzo Pellegrini Gabrielle Graffieti Vincenzo Lomonaco and Davide Maltoni. 2020. Latent replay for real-time continual learning. arXiv:1912.01100 [cs stat] (2020).","DOI":"10.1109\/IROS45743.2020.9341460"},{"key":"e_1_3_4_36_2","doi-asserted-by":"publisher","DOI":"10.1145\/3583683"},{"key":"e_1_3_4_37_2","unstructured":"Vikram Ramanathan. 2020. Online On-Device MCU Transfer Learning. Retrieved June 21 2023 from https:\/\/vikramramanathan.com\/files\/CS249R_Final_Report_Transfer_Learning.pdf."},{"key":"e_1_3_4_38_2","doi-asserted-by":"publisher","DOI":"10.1109\/JETCAS.2021.3121554"},{"key":"e_1_3_4_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.587"},{"key":"e_1_3_4_40_2","article-title":"TinyOL: TinyML with online-learning on microcontrollers","author":"Ren Haoyu","year":"2021","unstructured":"Haoyu Ren, Darko Anicic, and Thomas Runkler. 2021. TinyOL: TinyML with online-learning on microcontrollers. arXiv:2103.08295 [cs, eess] (2021). http:\/\/arxiv.org\/abs\/2103.08295.","journal-title":"arXiv:2103.08295 [cs, eess]"},{"key":"e_1_3_4_41_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-981-19-3391-2_2"},{"key":"e_1_3_4_42_2","doi-asserted-by":"publisher","DOI":"10.1038\/323533a0"},{"key":"e_1_3_4_43_2","first-page":"4167","volume-title":"Proceedings of the 2020 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP\u201920)","year":"2020","unstructured":"Simone Scardapane, Danilo Comminiello, Michele Scarpiniti, Enzo Baccarelli, and Aurello Uncini. 2020. Differentiable branching in deep networks for fast inference. In Proceedings of the 2020 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP\u201920). 4167\u20134171."},{"key":"e_1_3_4_44_2","first-page":"97","volume-title":"Proceedings of the 2021 IEEE SmartWorld Conference","year":"2021","unstructured":"Bharath Sudharsan, Piyush Yadav, John G. Breslin, and Muhammad Initzar Ali. 2021. Train++: An incremental ML model training algorithm to create self-learning IoT devices. In Proceedings of the 2021 IEEE SmartWorld Conference. 97\u2013106."},{"key":"e_1_3_4_45_2","unstructured":"Mingxing Tan and Quoc V. Le. 2019. EfficientNet: Rethinking model scaling for convolutional neural networks. arXiv:1905.11946 (2019). https:\/\/arxiv.org\/abs\/1905.11946."},{"key":"e_1_3_4_46_2","volume-title":"TinyML: Machine Learning with TensorFlow Lite on Arduino and Ultra-Low-Power Microcontrollers","author":"Warden Pete","year":"2020","unstructured":"Pete Warden and Daniel Situnayake. 2020. TinyML: Machine Learning with TensorFlow Lite on Arduino and Ultra-Low-Power Microcontrollers. O\u2019Reilly Media."},{"key":"e_1_3_4_47_2","doi-asserted-by":"publisher","DOI":"10.1023\/A:1018046501280"},{"key":"e_1_3_4_48_2","doi-asserted-by":"publisher","unstructured":"Zhengliang Zhu and Al.2020. A dataset of human motion status using IR-UWB through-wall radar. arXiv:2008.13598 (2020). 10.48550\/ARXIV.2008.13598","DOI":"10.48550\/ARXIV.2008.13598"}],"container-title":["ACM Transactions on Embedded Computing Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3604566","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3604566","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,17]],"date-time":"2025-06-17T16:47:20Z","timestamp":1750178840000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3604566"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,5,11]]},"references-count":47,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2024,5,31]]}},"alternative-id":["10.1145\/3604566"],"URL":"https:\/\/doi.org\/10.1145\/3604566","relation":{},"ISSN":["1539-9087","1558-3465"],"issn-type":[{"value":"1539-9087","type":"print"},{"value":"1558-3465","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,5,11]]},"assertion":[{"value":"2022-09-30","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-06-04","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2024-05-11","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}