{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,27]],"date-time":"2026-05-27T18:13:47Z","timestamp":1779905627241,"version":"3.53.1"},"reference-count":62,"publisher":"IOP Publishing","issue":"4","license":[{"start":{"date-parts":[[2023,11,24]],"date-time":"2023-11-24T00:00:00Z","timestamp":1700784000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2023,11,24]],"date-time":"2023-11-24T00:00:00Z","timestamp":1700784000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"DOI":"10.13039\/501100001804","name":"Canada Research Chair","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100001804","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Mach. Learn.: Sci. Technol."],"published-print":{"date-parts":[[2023,12,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Over the past decade, innovations in radiation and photonic detectors considerably improved their resolution, pixel density, sensitivity, and sampling rate, which all contribute to increased data generation rates. This huge data increases the amount of storage required, as well as the cabling between the source and the storage units. To overcome this problem, edge machine learning (EdgeML) proposes to move computation units near the detectors, utilizing machine learning (ML) models to emulate non-linear mathematical relationships between detector\u2019s output data. ML algorithms can be implemented in digital circuits, such as application-specific integrated circuits and field-programmable gate arrays, which support both parallelization and pipelining. EdgeML has both the benefits of edge computing and ML models to compress data near the detectors. This paper explores the currently available tool-flows designed to translate software ML algorithms to digital circuits near the edge. The main focus is on tool-flows that provide a diverse range of supported models, optimization techniques, and compression methods. We compare their accessibility, performance, and ease of use, and compare them for two high data-rate instrumentation applications: (1) CookieBox, and (2) billion-pixel camera.<\/jats:p>","DOI":"10.1088\/2632-2153\/ad0d12","type":"journal-article","created":{"date-parts":[[2023,11,16]],"date-time":"2023-11-16T22:25:15Z","timestamp":1700173515000},"page":"045035","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":4,"title":["Exploring machine learning to hardware implementations for large data rate x-ray instrumentation"],"prefix":"10.1088","volume":"4","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6582-8322","authenticated-orcid":true,"given":"Mohammad Mehdi","family":"Rahimifar","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-6301-8450","authenticated-orcid":true,"given":"Quentin","family":"Wingering","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Berthi\u00e9","family":"Gouin-Ferland","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hamza Ezzaoui","family":"Rahali","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Charles-\u00c9tienne","family":"Granger","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Audrey C","family":"Therrien","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"266","published-online":{"date-parts":[[2023,11,24]]},"reference":[{"key":"mlstad0d12bib1","doi-asserted-by":"publisher","first-page":"316","DOI":"10.3390\/jcm8030316","article-title":"The challenges of diagnostic imaging in the era of big data","volume":"8","author":"Aiello","year":"2019","journal-title":"J. Clin. Med."},{"key":"mlstad0d12bib2","doi-asserted-by":"publisher","DOI":"10.1088\/1402-4896\/ac6302","article-title":"An overview of the trigger system at the CMS experiment","volume":"97","author":"(CMS Collaboration)","year":"2022","journal-title":"Phys. Scr."},{"key":"mlstad0d12bib3","doi-asserted-by":"publisher","DOI":"10.1088\/1748-0221\/12\/05\/C05012","article-title":"Trigger systems of LHC experiments","volume":"12","author":"Jeitler","year":"2017","journal-title":"J. Instrum."},{"key":"mlstad0d12bib4","doi-asserted-by":"publisher","first-page":"62","DOI":"10.1016\/S0168-9002(01)01720-X","article-title":"Triggering at LHC experiments","volume":"478","author":"Smith","year":"2002","journal-title":"Nucl. Instrum. Methods Phys. Res. A"},{"key":"mlstad0d12bib5","article-title":"The ATLAS trigger and data acquisition upgrades for the high luminosity LHC (HL-LHC)","author":"Valente","year":"2019"},{"key":"mlstad0d12bib6","doi-asserted-by":"publisher","first-page":"215","DOI":"10.1038\/s41566-018-0107-6","article-title":"Attosecond time\u2013energy structure of x-ray free-electron laser pulses","volume":"12","author":"Hartmann","year":"2018","journal-title":"Nat. Photon."},{"key":"mlstad0d12bib7","first-page":"1","author":"Lin","year":"2020"},{"key":"mlstad0d12bib8","doi-asserted-by":"publisher","first-page":"219","DOI":"10.1016\/j.future.2019.02.050","article-title":"Edge computing: a survey","volume":"97","author":"Zada Khan","year":"2019","journal-title":"Future Gener. Comput. Syst."},{"key":"mlstad0d12bib9","doi-asserted-by":"publisher","first-page":"1537","DOI":"10.1109\/JPROC.2019.2920341","article-title":"A survey on edge computing systems and tools","volume":"107","author":"Liu","year":"2019","journal-title":"Proc. IEEE"},{"key":"mlstad0d12bib10","doi-asserted-by":"publisher","first-page":"1930","DOI":"10.1364\/AO.445798","article-title":"Potential of edge machine learning for instrumentation","volume":"61","author":"Therrien","year":"2022","journal-title":"Appl. Opt."},{"key":"mlstad0d12bib11","first-page":"pp 114","article-title":"Artificial intelligence (AI) chip technology review","author":"Li","year":"2019"},{"key":"mlstad0d12bib12","first-page":"pp 100","article-title":"A survey on convolutional neural network accelerators: GPU, FPGA and ASIC","author":"Hu","year":"2022"},{"key":"mlstad0d12bib13","first-page":"pp 1","article-title":"A systematic literature review of hardware neural networks","author":"Parra","year":"2018"},{"key":"mlstad0d12bib14","first-page":"pp 409","article-title":"Hardware accelerators for edge enabled machine learning","author":"Suresh","year":"2020"},{"key":"mlstad0d12bib15","doi-asserted-by":"publisher","DOI":"10.1088\/1748-0221\/17\/06\/C06009","article-title":"Machine learning on FPGA for event selection","volume":"17","author":"Furletov","year":"2022","journal-title":"J. Instrum."},{"key":"mlstad0d12bib16","doi-asserted-by":"publisher","DOI":"10.1088\/1748-0221\/16\/08\/P08016","article-title":"Nanosecond machine learning event classification with boosted decision trees in FPGA for high energy physics","volume":"16","author":"Hong","year":"2021","journal-title":"J. Instrum."},{"key":"mlstad0d12bib17","doi-asserted-by":"publisher","DOI":"10.1088\/1748-0221\/17\/07\/C07003","article-title":"High performance FPGA embedded system for machine learning based tracking and trigger in sPhenix and EIC","volume":"17","author":"Xuan","year":"2022","journal-title":"J. Instrum."},{"key":"mlstad0d12bib18","doi-asserted-by":"publisher","DOI":"10.1016\/j.nima.2022.167546","article-title":"Fast muon tracking with machine learning implemented in FPGA","volume":"1045","author":"Sun","year":"2023","journal-title":"Nucl. Instrum. Methods Phys. Res. A"},{"key":"mlstad0d12bib19","doi-asserted-by":"publisher","first-page":"1591","DOI":"10.1109\/TCAD.2015.2513673","article-title":"A survey and evaluation of FPGA high-level synthesis tools","volume":"35","author":"Nane","year":"2015","journal-title":"IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst."},{"key":"mlstad0d12bib20","doi-asserted-by":"publisher","first-page":"898","DOI":"10.1109\/TCAD.2018.2834439","article-title":"Are we there yet? A study on the state of high-level synthesis","volume":"38","author":"Lahti","year":"2018","journal-title":"IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst."},{"key":"mlstad0d12bib21","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3530775","article-title":"FPGA HLS today: successes, challenges and opportunities","volume":"15","author":"Cong","year":"2022","journal-title":"ACM Trans. Reconfigurable Technol. Syst."},{"key":"mlstad0d12bib22","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ac0ea1","article-title":"Fast convolutional neural networks on FPGAs with hls4ml","volume":"2","author":"Duarte","year":"2021","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstad0d12bib23","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3242897","article-title":"FINN-r: an end-to-end deep-learning framework for fast exploration of quantized neural networks","volume":"11","author":"Blott","year":"2018","journal-title":"ACM Trans. Reconfigurable Technol. Syst."},{"key":"mlstad0d12bib24","author":"Xilinx Vitis AI","year":"2022"},{"key":"mlstad0d12bib25","author":"TVM VTA","year":"2022"},{"key":"mlstad0d12bib26","author":"MATLAB DLP","year":"2022"},{"key":"mlstad0d12bib27","author":"OpenVINO 2023.2","year":"2022"},{"key":"mlstad0d12bib28","article-title":"OpenHLS: high-level synthesis for low-latency deep neural networks for experimental science","author":"Levental","year":"2023"},{"key":"mlstad0d12bib29","author":"NNgen: a fully-customizable hardware synthesis compiler for deep neural network","year":"2022"},{"key":"mlstad0d12bib30","first-page":"pp 741","article-title":"ScaleHLS: a new scalable high-level synthesis framework on multi-level intermediate representation","author":"Ye","year":"2022"},{"key":"mlstad0d12bib31","article-title":"CFU playground: full-stack open-source framework for tiny machine learning (tinyML) acceleration on FPGAs","author":"Prakash","year":"2022"},{"key":"mlstad0d12bib32","first-page":"pp 1","article-title":"VeriGOOD-ML: an open-source flow for automated ML hardware synthesis","author":"Esmaeilzadeh","year":"2021"},{"key":"mlstad0d12bib33","article-title":"DnnWeaver: from high-level deep network models to FPGA acceleration","author":"Sharma","year":"2016"},{"key":"mlstad0d12bib34","doi-asserted-by":"publisher","first-page":"2981","DOI":"10.3390\/s19132981","article-title":"Mapping neural networks to FPGA-based iot devices for ultra-low latency processing","volume":"19","author":"Wielgosz","year":"2019","journal-title":"Sensors"},{"key":"mlstad0d12bib35","first-page":"pp 40","article-title":"fpgaConvNet: a framework for mapping convolutional neural networks on FPGAs","author":"Venieris","year":"2016"},{"key":"mlstad0d12bib36","first-page":"pp 89","article-title":"FINN-L: library extensions and design trade-off analysis for variable precision lstm networks on FPGAs","author":"Rybalkin","year":"2018"},{"key":"mlstad0d12bib37","first-page":"pp 1","article-title":"LeFlow: enabling flexible FPGA high-level synthesis of tensorflow deep neural networks","author":"Noronha","year":"2018"},{"key":"mlstad0d12bib38","doi-asserted-by":"publisher","first-page":"196","DOI":"10.1016\/j.micpro.2018.03.007","article-title":"CaFPGA: an automatic generation model for CNN accelerator","volume":"60","author":"Xu","year":"2018","journal-title":"Microprocess. Microsyst."},{"key":"mlstad0d12bib39","first-page":"pp 40","article-title":"AutoDNNchip: an automated DNN chip predictor and builder for both FPGAs and ASICs","author":"Xu","year":"2020"},{"key":"mlstad0d12bib40","first-page":"pp 152","article-title":"FP-DNN: an automated framework for mapping deep neural networks onto FPGAs with RTL-HLS hybrid templates","author":"Guan","year":"2017"},{"key":"mlstad0d12bib41","first-page":"pp 1","article-title":"Snowflake: an efficient hardware accelerator for convolutional neural networks","author":"Gokhale","year":"2017"},{"key":"mlstad0d12bib42","first-page":"pp 117","article-title":"A framework for generating high throughput CNN implementations on FPGAs","author":"Zeng","year":"2018"},{"key":"mlstad0d12bib43","first-page":"pp 69","article-title":"A holistic approach for optimizing DSP block utilization of a CNN implementation on FPGA","author":"Abdelouahab","year":"2016"},{"key":"mlstad0d12bib44","doi-asserted-by":"publisher","first-page":"35","DOI":"10.1109\/TCAD.2017.2705069","article-title":"Angel-Eye: a complete design flow for mapping CNN onto embedded FPGA","volume":"37","author":"Guo","year":"2017","journal-title":"IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst."},{"key":"mlstad0d12bib45","doi-asserted-by":"publisher","first-page":"2072","DOI":"10.1109\/TCAD.2017.2785257","article-title":"Caffeine: toward uniformed representation and acceleration for deep convolutional neural networks","volume":"38","author":"Zhang","year":"2018","journal-title":"IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst."},{"key":"mlstad0d12bib46","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/acc0d7","article-title":"Ultra-low latency recurrent neural network inference on FPGAs for physics applications with hls4ml","volume":"4","author":"Khoda","year":"2023","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstad0d12bib47","doi-asserted-by":"publisher","DOI":"10.1088\/1748-0221\/13\/07\/P07027","article-title":"Fast inference of deep neural networks in FPGAs for particle physics","volume":"13","author":"Duarte","year":"2018","journal-title":"J. Instrum."},{"key":"mlstad0d12bib48","first-page":"pp 10","article-title":"A safari through FPGA-based neural network compilation and design automation flows","author":"Plagwitz","year":"2021"},{"key":"mlstad0d12bib49","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ac9cb5","article-title":"Real-time semantic segmentation on FPGAs for autonomous vehicles with hls4ml","volume":"3","author":"Ghielmetti","year":"2022","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstad0d12bib50","article-title":"Open-source FPGA-ML codesign for the MLPerf Tiny Benchmark","author":"Borras","year":"2022"},{"key":"mlstad0d12bib51","author":"Xilinx Vitis AI Model Zoo","year":"2023"},{"key":"mlstad0d12bib52","author":"HDLCoder","year":"2023"},{"key":"mlstad0d12bib53","article-title":"Multi-level intermediate representation overview","author":"MLIR","year":"2023"},{"key":"mlstad0d12bib54","author":"UNSW-NB15 data set","year":"2023"},{"key":"mlstad0d12bib55","article-title":"The Street View House Numbers (SVHN) Dataset","year":"2023"},{"key":"mlstad0d12bib56","author":"HLS4ML_Additional_Boards","year":"2023"},{"key":"mlstad0d12bib57","doi-asserted-by":"publisher","DOI":"10.1063\/5.0043013","article-title":"Billion-pixel x-ray camera (BiPC-X)","volume":"92","author":"Wang","year":"2021","journal-title":"Rev. Sci. Instrum."},{"key":"mlstad0d12bib58","first-page":"pp 1","article-title":"Machine learning at the edge for ultra high rate detectors","author":"Corbeil Therrien","year":"2019"},{"key":"mlstad0d12bib59","first-page":"pp 1","article-title":"Combining optimized quantization and machine learning for real-time data reduction at the edge","author":"Gouin-Ferland","year":"2022"},{"key":"mlstad0d12bib60","doi-asserted-by":"publisher","first-page":"223","DOI":"10.1016\/j.nima.2019.06.011","article-title":"Ultrafast inorganic scintillator-based front imager for gigahertz hard x-ray imaging","volume":"940","author":"Hu","year":"2019","journal-title":"Nucl. Instrum. Methods Phys. Res. A"},{"key":"mlstad0d12bib61","doi-asserted-by":"publisher","DOI":"10.1016\/j.nima.2023.168829","article-title":"Efficient compression at the edge for real-time data acquisition in a billion-pixel x-ray camera","volume":"1058","author":"Ezzaoui Rahali","year":"2024","journal-title":"Nucl. Instrum. Methods Phys. Res. A"},{"key":"mlstad0d12bib62","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/aba042","article-title":"Compressing deep neural networks on FPGAs to binary and ternary precision with hls4ml","volume":"2","author":"Ngadiuba","year":"2020","journal-title":"Mach. Learn.: Sci. Technol."}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,11,24]],"date-time":"2023-11-24T06:42:44Z","timestamp":1700808164000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad0d12"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,11,24]]},"references-count":62,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2023,11,24]]},"published-print":{"date-parts":[[2023,12,1]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/ad0d12","relation":{},"ISSN":["2632-2153"],"issn-type":[{"value":"2632-2153","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,11,24]]},"assertion":[{"value":"Exploring machine learning to hardware implementations for large data rate x-ray instrumentation","name":"article_title","label":"Article Title"},{"value":"Machine Learning: Science and Technology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2023 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2023-07-17","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2023-11-15","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2023-11-24","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}