{"id":13161,"date":"2025-04-11T13:46:24","date_gmt":"2025-04-11T04:46:24","guid":{"rendered":"https:\/\/blogs.mathworks.com\/japan-community\/?p=13161"},"modified":"2025-04-11T13:46:24","modified_gmt":"2025-04-11T04:46:24","slug":"edge-ai-foundations-austin-2025-jp","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/japan-community\/2025\/04\/11\/edge-ai-foundations-austin-2025-jp\/","title":{"rendered":"Edge AI Foundation\u2019s Austin 2025 \u4fbf\u308a"},"content":{"rendered":"<div style=\"margin: 2px 10px 9px 4px; padding: 0px; line-height: 21px; min-height: 0px; white-space: pre-wrap; color: #212121; font-family: Helvetica, Arial, sans-serif; font-style: normal; font-size: 14px; font-weight: 400; text-align: left;\">\u203b\u3053\u306e\u6295\u7a3f\u306f 2025 \u5e74 3 \u6708 24 \u65e5\u306b <a href=\"https:\/\/blogs.mathworks.com\/deep-learning\/2025\/03\/24\/highlights-from-the-edge-ai-foundations-austin-2025-conference\/\">Artificial Intelligence \u3078 \u6295\u7a3f<\/a>\u3055\u308c\u305f\u3082\u306e\u306e\u6284\u8a33\u3067\u3059\u3002<\/div>\n<hr \/>\n<p><em>The following blog post is from <\/em><a href=\"https:\/\/www.linkedin.com\/in\/jack-ferrari\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\"><em>Jack Ferrari<\/em><\/a><em>, Edge AI Product Manager, <\/em><a href=\"https:\/\/www.linkedin.com\/in\/joedsanford\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\"><em>Joe Sanford<\/em><\/a><em>, Senior Application Engineer, and <\/em><a href=\"https:\/\/www.linkedin.com\/in\/reed-axman-43287a13b\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\"><em>Reed Axman<\/em><\/a><em>, Strategic Partner Manager. <\/em><\/p>\n<p>\u30c6\u30ad\u30b5\u30b9\u5dde\u30aa\u30fc\u30b9\u30c6\u30a3\u30f3\u3067\u958b\u50ac\u3055\u308c\u305f Edge AI Foundation \u306e\u30ab\u30f3\u30d5\u30a1\u30ec\u30f3\u30b9\u306b\u53c2\u52a0\u3057\u307e\u3057\u305f\u3002\u3053\u306e\u30a4\u30d9\u30f3\u30c8\u306f\u3001\u5b66\u8005\u3001\u7814\u7a76\u8005\u3001\u30d3\u30b8\u30cd\u30b9\u30ea\u30fc\u30c0\u30fc\u3001\u30a8\u30f3\u30b8\u30cb\u30a2\u3092\u96c6\u3081\u3001AI \u30e2\u30c7\u30eb\u3092\u5b9f\u9a13\u5ba4\u304b\u3089\u5b9f\u969b\u306e\u30a8\u30c3\u30b8\u30c7\u30d0\u30a4\u30b9\u3078\u3068\u5c55\u958b\u3059\u308b\u3053\u3068\u306b\u7126\u70b9\u3092\u5f53\u3066\u3066\u3044\u307e\u3059\u30023 \u65e5\u9593\u306b\u308f\u305f\u308b\u9b45\u529b\u7684\u306a\u30a4\u30d9\u30f3\u30c8\u3067\u3001MathWorks \u306e\u30c1\u30fc\u30e0\u306f\u591a\u304f\u306e\u53c2\u52a0\u8005\u30fb\u8b1b\u6f14\u8005\u3068\u4ea4\u6d41\u3057\u307e\u3057\u305f\u3002<\/p>\n<p>Edge AI \u306e\u30b9\u30dd\u30f3\u30b5\u30fc\u3068\u3057\u3066 MathWorks \u306f\u30c7\u30e2\u30d6\u30fc\u30b9\u3092\u8a2d\u3051\u30012 \u3064\u306e\u30e9\u30a4\u30d6\u30c7\u30e2\u3092\u884c\u3044\u3001\u3055\u3089\u306b MATLAB \u3068 Simulink \u3092\u4f7f\u7528\u3057\u3066\u30e2\u30fc\u30bf\u30fc\u5236\u5fa1\u306e\u305f\u3081\u306e AI \u30e2\u30c7\u30eb\u3092\u958b\u767a\u3059\u308b\u65b9\u6cd5\u3092\u7d39\u4ecb\u3059\u308b\u30cf\u30f3\u30ba\u30aa\u30f3\u30ef\u30fc\u30af\u30b7\u30e7\u30c3\u30d7\u3092\u5b9f\u65bd\u3057\u307e\u3057\u305f\u3002\u5927\u304d\u304f\u6210\u9577\u3057\u3066\u3044\u308b Edge AI\u00a0 \u30b3\u30df\u30e5\u30cb\u30c6\u30a3\u304b\u3089\u5b66\u3073\u3001\u60c5\u5831\u3092\u5171\u6709\u3057\u3001\u8ca2\u732e\u3082\u3055\u305b\u3066\u3082\u3089\u3046\u7d20\u6674\u3089\u3057\u3044\u6a5f\u4f1a\u3067\u3057\u305f\u3002<\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17099 size-full\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/03\/figure1.jpg\" alt=\"\" width=\"1156\" height=\"868\" \/><\/p>\n<p><strong>Figure 1:<\/strong> Edge AI Foundation 2025 \u3067\u306e MathWorks \u30d6\u30fc\u30b9<\/p>\n<h6><\/h6>\n<p style=\"font-size: 20px; color: #c04c0b;\"><strong>MathWorks \u30d6\u30fc\u30b9\u3067\u306e\u30c7\u30e2\u5c55\u793a<\/strong><\/p>\n<p>AI \u30e2\u30c7\u30eb\u306e\u958b\u767a\u3084\u30a8\u30c3\u30b8\u30c7\u30d0\u30a4\u30b9\u3078\u306e\u5c55\u958b\u306e\u30d7\u30ed\u30bb\u30b9\u3092 \u3001MATLAB \u3068 Simulink \u3092\u4f7f\u3063\u3066\u52a0\u901f\u3059\u308b\u65b9\u6cd5\u3092\u7d39\u4ecb\u3059\u308b\u30c7\u30e2\u3092 2 \u3064\u767a\u8868\u3057\u307e\u3057\u305f\u3002<\/p>\n<p style=\"font-size: 18px;\"><strong>\u30d0\u30c3\u30c6\u30ea\u30fc\u5145\u96fb\u72b6\u614b\u306e\u4eee\u60f3\u30bb\u30f3\u30b5\u30fc<\/strong><\/p>\n<p>\u3053\u306e\u30c7\u30e2\u3067\u306f\u3001<a href=\"https:\/\/blogs.mathworks.com\/deep-learning\/2023\/08\/11\/virtual-sensors-with-ai-and-model-based-design\/\">AI \u30d9\u30fc\u30b9\u306e\u4eee\u60f3\u30bb\u30f3\u30b5\u30fc<\/a>\u304c\u3001\u96fb\u5727\u3001\u96fb\u6d41\u3001\u6e29\u5ea6\u306e\u6e2c\u5b9a\u5024\u304b\u3089\u30ea\u30c1\u30a6\u30e0\u30a4\u30aa\u30f3\u30d0\u30c3\u30c6\u30ea\u30fc\u306e<a href=\"https:\/\/www.mathworks.com\/discovery\/battery-state-of-charge.html\">\u5145\u96fb\u72b6\u614b\uff08SoC\uff09<\/a>\u3092\u30ea\u30a2\u30eb\u30bf\u30a4\u30e0\u4e88\u6e2c\u3059\u308b\u69d8\u5b50\u3092\u793a\u3057\u307e\u3057\u305f\u3002\u30d0\u30c3\u30c6\u30ea\u30fc\u306e SoC \u63a8\u5b9a\u306f\u3001\u591a\u304f\u306e\u30d0\u30c3\u30c6\u30ea\u30fc\u99c6\u52d5\u30b7\u30b9\u30c6\u30e0\u306b\u3068\u3063\u3066\u975e\u5e38\u306b\u91cd\u8981\u3067\u3042\u308a\u3001\u76f4\u63a5\u6e2c\u5b9a\u3059\u308b\u3053\u3068\u304c\u96e3\u3057\u3044\u5834\u5408\u3084\u4e0d\u53ef\u80fd\u306a\u5834\u5408\u304c\u3042\u308a\u307e\u3059\u3002\u30cb\u30e5\u30fc\u30e9\u30eb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u3088\u3046\u306a\u30c7\u30fc\u30bf\u99c6\u52d5\u578b\u30e2\u30c7\u30eb\u3092\u4f7f\u7528\u3059\u308b\u3053\u3068\u3067\u3001\u95a2\u9023\u3059\u308b\u6e2c\u5b9a\u5024\u306b\u57fa\u3065\u3044\u3066\u6b63\u78ba\u306a\u4e88\u6e2c\u304c\u53ef\u80fd\u306b\u306a\u308a\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u30c7\u30e2\u3067\u306f\u3001Deep Learning Toolbox \u3092\u4f7f\u7528\u3057\u3066\u3001\u5b9f\u969b\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3067\u7c21\u5358\u306a LSTM \u30e2\u30c7\u30eb\u3092\u5b66\u7fd2\u3057\u3001\u30ea\u30bd\u30fc\u30b9\u5236\u7d04\u306e\u3042\u308b\u7d44\u307f\u8fbc\u307f\u30c7\u30d0\u30a4\u30b9\u3078\u306e\u5b9f\u88c5\u3059\u308b\u305f\u3081\u306b <a href=\"https:\/\/www.mathworks.com\/company\/technical-articles\/compressing-neural-networks-using-network-projection.html\">neural network projection <\/a>(\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u6295\u5f71)\u3092\u7528\u3044\u307e\u3057\u305f\u3002\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u6295\u5f71\u306b\u3088\u308a LSTM \u30e2\u30c7\u30eb\u306e\u5b66\u7fd2\u53ef\u80fd\u306a\u30d1\u30e9\u30e1\u30fc\u30bf\u6570\u306f\u7d04 4 \u4e07\u304b\u3089\uff13\u5343 \u3078\u3068 92% \u524a\u6e1b\u3055\u308c\u3001\u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3067\u306e\u7cbe\u5ea6\u4f4e\u4e0b\u306f\u3042\u308a\u307e\u305b\u3093\u3067\u3057\u305f\u3002\u5727\u7e2e\u5f8c\u306e\u30e2\u30c7\u30eb\u30d1\u30e9\u30e1\u30fc\u30bf\u306e\u30b5\u30a4\u30ba\u306f 25.7 kB \u3067\u3057\u305f\u3002<\/p>\n<div class=\"max-h-full overflow-x-hidden\">\n<div>\n<div class=\"group px-4 border-b border-black\/10 bg-gray-50 text-gray-800 dark:border-gray-900\/50 dark:bg-[var(--dark-gray-background-color)] dark:text-gray-100\">\n<div class=\"relative m-auto flex gap-3 p-3 text-base md:max-w-2xl md:gap-2 md:py-3 lg:max-w-2xl lg:px-0 xl:max-w-3xl sm:mx-auto\">\n<div class=\"mt-[-2px] w-full prose dark:prose-invert \">\n<div class=\"flex w-full flex-col\">\n<div class=\"prose max-w-none dark:prose-invert\">\n<p><a href=\"https:\/\/www.mathworks.com\/help\/coder\/deep-learning-with-matlab-coder.html\">MATLAB Coder<\/a> \u3092\u4f7f\u7528\u3057\u3066 LSTM \u30e2\u30c7\u30eb\u304b\u3089\u30b9\u30bf\u30f3\u30c9\u30a2\u30ed\u30f3\u306e C \u30b3\u30fc\u30c9\u3092\u751f\u6210\u3057\u3001\u30c7\u30e2\u30ad\u30c3\u30c8\u306e<a href=\"https:\/\/github.com\/Infineon\/IMR_PSoC6_BMS?tab=readme-ov-file\" class=\"external\" rel=\"nofollow\" target=\"_blank\">\u65e2\u5b58\u306e BMS \u30d5\u30a1\u30fc\u30e0\u30a6\u30a7\u30a2<\/a>\u306b\u7d71\u5408\u3057\u307e\u3057\u305f\u3002\u3053\u306e\u30c7\u30e2\u3067\u4f7f\u7528\u3055\u308c\u305f\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u306f\u3001<a href=\"https:\/\/www.infineon.com\/cms\/en\/applications\/robotics\/development-platform\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\">Infineon Mobile Robot<\/a> \u3067\u4f7f\u7528\u3055\u308c\u3066\u3044\u308b BMS \u30b3\u30f3\u30c8\u30ed\u30fc\u30e9\u30fc\u306b\u57fa\u3065\u3044\u3066\u3044\u307e\u3059\u3002<\/p>\n<\/div>\n<div class=\"flex justify-start mb-6 pt-2\"><\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17102 size-full\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/03\/figure2.png\" alt=\"\" width=\"1466\" height=\"771\" \/><\/p>\n<p><strong>Figure 2:<\/strong> \u30d0\u30c3\u30c6\u30ea\u30fc\u306e SoC \u63a8\u5b9a\u4eee\u60f3\u30bb\u30f3\u30b5\u30fc\u30c7\u30e2<\/p>\n<h6><\/h6>\n<p style=\"font-size: 18px;\"><strong>\u30b9\u30de\u30fc\u30c8\u30b9\u30d4\u30fc\u30ab\u30fc\u306e\u30a6\u30a7\u30a4\u30af\u30ef\u30fc\u30c9\uff08wake word\uff09\u691c\u51fa<\/strong><\/p>\n<p>\u3053\u306e\u30c7\u30e2\u3067\u306f\u3001Simulink \u3068 <a href=\"https:\/\/www.mathworks.com\/hardware-support\/qualcomm-hexagon.html\">Embedded Coder Support Package for Qualcomm Hexagon Processors<\/a> \u3092\u4f7f\u7528\u3057\u3066\u3001Qualcomm \u306e Hexagon \u4e0a\u3067\u97f3\u58f0\u30b3\u30de\u30f3\u30c9\u8a8d\u8b58\u306e\u305f\u3081\u306e<a href=\"https:\/\/www.mathworks.com\/discovery\/convolutional-neural-network.html\">\u7573\u307f\u8fbc\u307f\u30cb\u30e5\u30fc\u30e9\u30eb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af<\/a>\uff08CNN\uff09\u306e\u5b9f\u88c5\u3092\u884c\u3044\u307e\u3057\u305f\u3002\u30c7\u30e2\u306f\u30b9\u30de\u30fc\u30c8\u30b9\u30d4\u30fc\u30ab\u30fc\u306e\u30a6\u30a7\u30a4\u30af\u30ef\u30fc\u30c9\u691c\u51fa\u3092\u5bfe\u8c61\u3068\u3057\u3066\u3044\u307e\u3059\u3002Hexagon \u5411\u3051\u306e\u6700\u9069\u5316\u3055\u308c\u305f\u30b3\u30fc\u30c9\u751f\u6210\u306e\u5229\u70b9\u3092\u5f37\u8abf\u3057\u3001\u3053\u306e\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u30b5\u30dd\u30fc\u30c8\u30d1\u30c3\u30b1\u30fc\u30b8\u304cQualcomm \u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u306e\u4f7f\u7528\u3092\u3069\u308c\u3060\u3051\u7c21\u7d20\u5316\u3059\u308b\u304b\u3092\u7d39\u4ecb\u3057\u307e\u3057\u305f\u3002Qualcomm \u306e\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u30c4\u30fc\u30eb\u306b\u95a2\u3059\u308b\u5c02\u9580\u77e5\u8b58\u3092\u6301\u305f\u305a\u306b\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u3092\u6d3b\u7528\u3067\u304d\u3001AI \u30e2\u30c7\u30eb\u306eeNPU \u3078\u306e\u5c55\u958b\u3092 5 \u5206\u4ee5\u5185\u3067\u53d6\u308a\u304b\u304b\u308b\u3053\u3068\u304c\u53ef\u80fd\u306b\u306a\u308a\u307e\u3059\u3002<\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17141 size-full\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/03\/SL_model.jpeg\" alt=\"\" width=\"800\" height=\"159\" \/><\/p>\n<p><strong>Figure 3:<\/strong> \u30b9\u30de\u30fc\u30c8\u30b9\u30d4\u30fc\u30ab\u30fc\u30c7\u30e2\u3067\u4f7f\u7528\u3057\u305f Simulink \u30e2\u30c7\u30eb<\/p>\n<h6><\/h6>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17108 size-full\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/03\/figure4.jpg\" alt=\"\" width=\"998\" height=\"998\" \/><\/p>\n<p><strong>Figure 4:<\/strong> Qualcomm Hexagon eNPU \u306b\u5b9f\u88c5\u3057\u305f\u30a6\u30a7\u30a4\u30af\u30ef\u30fc\u30c9\u691c\u51fa\u30c7\u30e2<\/p>\n<p>&nbsp;<\/p>\n<p>\u3053\u308c\u3089\u306e\u30cf\u30f3\u30ba\u30aa\u30f3\u30c7\u30e2\u306b\u52a0\u3048\u3066\u3001<a href=\"https:\/\/www.mathworks.com\/help\/deeplearning\/ref\/deepnetworkdesigner-app.html\">Deep Network Designer <\/a>\u3001<a href=\"https:\/\/www.mathworks.com\/help\/deeplearning\/ug\/analyze-and-compress-1d-convolutional-neural-network.html#AnalyzeAndCompress1DConvolutionalNeuralNetworkExample-2\">Analyze for Compression<\/a>\u3001<a href=\"https:\/\/www.mathworks.com\/help\/predmaint\/ref\/diagnosticfeaturedesigner-app.html\">Diagnostic Feature Designer<\/a>\u3001<a href=\"https:\/\/www.mathworks.com\/help\/stats\/classificationlearner-app.html\">Classification Learner<\/a>\u3001<a href=\"https:\/\/www.mathworks.com\/help\/deeplearning\/ref\/experimentmanager-app.html\">Experiment Manager<\/a>\u3001\u304a\u3088\u3073<a href=\"https:\/\/www.mathworks.com\/help\/signal\/ref\/signallabeler-app.html\">Signal Labeler<\/a>\u00a0\u3084 <a href=\"https:\/\/www.mathworks.com\/help\/vision\/ref\/imagelabeler-app.html\">Image Labeler<\/a> \u306a\u3069\u306e\u30a2\u30d7\u30ea\u306e\u5373\u8208\u30c7\u30e2\u3082\u884c\u3044\u307e\u3057\u305f\u3002<\/p>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17111 size-full\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/03\/figure5.jpg\" alt=\"\" width=\"1263\" height=\"843\" \/><\/p>\n<p><strong>Figure 5:<\/strong>\u00a0MathWorks \u30d6\u30fc\u30b9\u3067\u306e\u4e00\u5e55<\/p>\n<h6><\/h6>\n<p>&nbsp;<\/p>\n<h6><\/h6>\n<p style=\"font-size: 20px; color: #c04c0b;\"><strong>STMicroelectronics \u3068\u5171\u540c\u3067\u884c\u3063\u305f\u30ef\u30fc\u30af\u30b7\u30e7\u30c3\u30d7<\/strong><\/p>\n<p>&#8220;<em>Rev it up: Deploy Tiny Neural Network to Boost Embedded Field-Oriented Controls of Electrical Drives&#8221;<\/em>\u3068\u984c\u3057\u305f\u30ef\u30fc\u30af\u30b7\u30e7\u30c3\u30d7\u3067\u306f\u3001\u53c2\u52a0\u8005\u306b\u30e2\u30fc\u30bf\u30fc\u5236\u5fa1\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u3067\u4f7f\u7528\u3055\u308c\u308b\u53e4\u5178\u7684\u306a\u6bd4\u4f8b\u7a4d\u5206\uff08PI\uff09\u30b3\u30f3\u30c8\u30ed\u30fc\u30e9\u30fc\u3092\u6539\u5584\u3059\u308b\u305f\u3081\u306b\u3001\u5c0f\u578b\u306e\u30cb\u30e5\u30fc\u30e9\u30eb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u3092\u8a2d\u8a08\u3001\u6700\u9069\u5316\u3001\u304a\u3088\u3073\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3059\u308b\u8ab2\u984c\u3092\u5b9f\u65bd\u3057\u307e\u3057\u305f\u3002\u53c2\u52a0\u8005\u306f\u3001Simulink \u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\u304b\u3089\u5408\u6210\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u30c7\u30fc\u30bf\u3092\u751f\u6210\u3057\u3001Deep Learning Toolbox \u3067\u3055\u307e\u3056\u307e\u306a\u6df1\u5c64\u5b66\u7fd2\u30e2\u30c7\u30eb\u3092\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u304a\u3088\u3073\u5727\u7e2e\u3057\u3001<a href=\"https:\/\/stm32ai.st.com\/st-edge-ai-developer-cloud\/\" class=\"external\" rel=\"nofollow\" target=\"_blank\">ST Edge AI Developer Cloud<\/a> \u304b\u3089\u30a2\u30af\u30bb\u30b9\u53ef\u80fd\u306a\u30ea\u30e2\u30fc\u30c8\u30dc\u30fc\u30c9\u30d5\u30a1\u30fc\u30e0\u4e0a\u3067\u30e2\u30c7\u30eb\u306e\u5b9f\u884c\u6642\u6027\u80fd\u3092\u8a55\u4fa1\u3057\u307e\u3057\u305f\u3002<\/p>\n<p>\u3053\u306e\u30ef\u30fc\u30af\u30b7\u30e7\u30c3\u30d7\u3092\u958b\u50ac\u3059\u308b\u6a5f\u4f1a\u3092\u4e0e\u3048\u3066\u304f\u308c\u305f Edge AI Foundation \u3068\u3001\u5171\u306b\u767a\u8868\u3057\u3066\u304f\u308c\u305f STMicroelectronics \u306e Dr. Danilo Pau \u306b\u5fc3\u304b\u3089\u611f\u8b1d\u3057\u3066\u3044\u307e\u3059\u3002\u53c2\u52a0\u8005\u304b\u3089\u306e\u79f0\u8cdb\u3068\u6539\u5584\u63d0\u6848\u3092\u805e\u304f\u3053\u3068\u304c\u3067\u304d\u3066\u5b09\u3057\u304b\u3063\u305f\u3067\u3059\u3002\u591a\u304f\u306e\u53c2\u52a0\u8005\u306f\u300190 \u5206\u306e\u30a4\u30d9\u30f3\u30c8\u306e\u4e2d\u3067\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u30c7\u30fc\u30bf\u306e\u751f\u6210\u304b\u3089 STM32 \u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u3078\u306e\u5c55\u958b\u3068\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u307e\u3067\u3001\u30a8\u30f3\u30c9\u30c4\u30fc\u30a8\u30f3\u30c9\u306e\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u3092\u5b8c\u4e86\u3055\u305b\u3066\u3044\u307e\u3057\u305f\u3002\u3053\u306e\u30ef\u30fc\u30af\u30b7\u30e7\u30c3\u30d7\u306b\u3064\u3044\u3066\u8a73\u3057\u304f\u306f\u3001<a href=\"https:\/\/content.mathworks.com\/viewer\/67d32b8ade6f00f5710896ee\">\u30ef\u30fc\u30af\u30b7\u30e7\u30c3\u30d7\u306e\u6982\u8981<\/a>\u3068<a href=\"https:\/\/github.com\/Brenda-MW\/RevUp-EAIF-AUSTIN-2025\/tree\/main\" class=\"external\" rel=\"nofollow\" target=\"_blank\">\u30ea\u30dd\u30b8\u30c8\u30ea<\/a>\u3092\u3054\u89a7\u304f\u3060\u3055\u3044\u3002<\/p>\n<h6><\/h6>\n<p><img decoding=\"async\" loading=\"lazy\" class=\"alignnone wp-image-17114 size-full\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/03\/figure6.jpg\" alt=\"\" width=\"1256\" height=\"816\" \/><\/p>\n<p><strong>Figure 6:<\/strong> Rev it up: Deploy Tiny Neural Network to Boost Embedded Field-Oriented Controls of Electrical Drives \u30ef\u30fc\u30af\u30b7\u30e7\u30c3\u30d7\u306e\u69d8\u5b50<\/p>\n<h6><\/h6>\n<h6><\/h6>\n<p style=\"font-size: 20px; color: #c04c0b;\"><strong>\u307e\u3068\u3081\u3068\u696d\u754c\u30c8\u30ec\u30f3\u30c9<\/strong><\/p>\n<p>\u3053\u306e\u30ab\u30f3\u30d5\u30a1\u30ec\u30f3\u30b9\u3067\u306f\u3001\u6025\u901f\u306b\u9032\u5316\u3059\u308b\u30a8\u30c3\u30b8 AI \u306e\u4e16\u754c\u306b\u95a2\u3059\u308b\u6d1e\u5bdf\u306b\u3042\u3075\u308c\u3066\u3044\u307e\u3057\u305f\u3002\u30aa\u30fc\u30b9\u30c6\u30a3\u30f3\u3067\u306e 3 \u65e5\u9593\u3067\u7279\u306b\u6ce8\u76ee\u3092\u96c6\u3081\u305f\u306e\u306f\u6b21\u306e\u30c8\u30ec\u30f3\u30c9\u3067\u3059\u3002<\/p>\n<ul>\n<li><strong>\u7269\u7406 AI \u3068\u30ed\u30dc\u30c6\u30a3\u30af\u30b9\u3078\u306e\u95a2\u5fc3\u306e\u9ad8\u307e\u308a<\/strong>\uff1a\u30ed\u30dc\u30c3\u30c8\u3084\u7269\u7406\u74b0\u5883\u3068\u76f8\u4e92\u4f5c\u7528\u3059\u308b\u4ed6\u306e\u81ea\u5f8b\u7684\u306a\u30b7\u30b9\u30c6\u30e0\u306bAI\u3092\u7d71\u5408\u3059\u308b\u3053\u3068\u3078\u306e\u95a2\u5fc3\u304c\u9855\u8457\u306b\u9ad8\u307e\u3063\u3066\u3044\u307e\u3059\u3002\u30b7\u30b9\u30c6\u30e0\u304c\u3088\u308a\u8907\u96d1\u306b\u306a\u308b\u306b\u3064\u308c\u3066\u3001\u5305\u62ec\u7684\u306a\u8a2d\u8a08\u3068\u30c6\u30b9\u30c8\u30d7\u30ed\u30bb\u30b9\u3092\u901a\u3058\u3066\u305d\u306e\u4fe1\u983c\u6027\u3068\u52b9\u679c\u3092\u78ba\u4fdd\u3059\u308b\u3053\u3068\u304c\u91cd\u8981\u306b\u306a\u308a\u307e\u3059\u3002\u3053\u308c\u3089\u306e\u30b7\u30b9\u30c6\u30e0\u3092\u8fc5\u901f\u306b\u30d7\u30ed\u30c8\u30bf\u30a4\u30d7\u5316\u3057\u3001\u53b3\u5bc6\u306b\u30c6\u30b9\u30c8\u3059\u308b\u80fd\u529b\u304c\u3001\u6210\u529f\u3057\u305f\u5c55\u958b\u3068\u5b89\u5168\u306a\u904b\u7528\u306e\u9375\u3068\u306a\u308b\u3067\u3057\u3087\u3046\u3002<\/li>\n<li><strong>\u5c02\u7528\u30cb\u30e5\u30fc\u30e9\u30eb\u30d7\u30ed\u30bb\u30c3\u30b7\u30f3\u30b0\u30e6\u30cb\u30c3\u30c8\uff08NPU\uff09\u306e\u62e1\u5927<\/strong>\uff1a\u304b\u3064\u3066\u306f PC \u3084\u30b9\u30de\u30fc\u30c8\u30d5\u30a9\u30f3\u306b\u9650\u3089\u308c\u3066\u3044\u305f NPU \u304c\u3001\u3088\u308a\u591a\u304f\u306e\u30de\u30a4\u30af\u30ed\u30b3\u30f3\u30c8\u30ed\u30fc\u30e9\u30fc\u3084 IoT \u30c7\u30d0\u30a4\u30b9\u306b\u9032\u51fa\u3057\u7d9a\u3051\u3066\u3044\u307e\u3059\u3002\u3053\u306e\u5909\u5316\u306f\u3001\u5f93\u6765\u306e CPU \u3068\u6bd4\u8f03\u3057\u3066\u3001\u96fb\u529b\u52b9\u7387\u3068\u6027\u80fd\u306e\u5411\u4e0a\u306e\u53ef\u80fd\u6027\u3092\u793a\u3057\u3066\u304a\u308a\u3001\u30a8\u30c3\u30b8\u3067\u306e\u30c7\u30fc\u30bf\u51e6\u7406\u306b\u304a\u3051\u308b\u91cd\u8981\u306a\u9032\u5316\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002MATLAB \u30e6\u30fc\u30b6\u30fc\u3067 NPU \u306b AI \u30e2\u30c7\u30eb\u3092\u5c55\u958b\u3059\u308b\u3053\u3068\u306b\u8208\u5473\u304c\u3042\u308b\u65b9\u306f\u3001<a href=\"https:\/\/www.mathworks.com\/solutions\/contact-the-expert.html?tfa_5=Embedded+AI\">\u305c\u3072\u304a\u77e5\u3089\u305b\u304f\u3060\u3055\u3044<\/a>\uff01\u7686\u3055\u3093\u306e\u6d1e\u5bdf\u3068\u7d4c\u9a13\u304c\u3001\u8ffd\u52a0\u306e NPU \u30bf\u30fc\u30b2\u30c3\u30c8\u3078\u306e\u30b5\u30dd\u30fc\u30c8\u3092\u691c\u8a0e\u3059\u308b\u4e2d\u3067\u8cb4\u91cd\u306a\u60c5\u5831\u306b\u306a\u308a\u307e\u3059\u3002<\/li>\n<li><strong>\u30cb\u30e5\u30fc\u30e9\u30eb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u5727\u7e2e\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u306e\u9032\u5c55<\/strong>\uff1a\u65b0\u3057\u3044\u5727\u7e2e\u6280\u8853\u306f AI \u30e2\u30c7\u30eb\u306e\u30b5\u30a4\u30ba\u3068\u8a08\u7b97\u8981\u6c42\u3092\u5927\u5e45\u306b\u524a\u6e1b\u3057\u3001\u307e\u3059\u307e\u3059\u8907\u96d1\u306a\u30e2\u30c7\u30eb\u3092\u30a8\u30c3\u30b8\u30c7\u30d0\u30a4\u30b9\u306b\u5c55\u958b\u3059\u308b\u3053\u3068\u3092\u53ef\u80fd\u306b\u3057\u3066\u3044\u307e\u3059\u3002\u30ab\u30f3\u30d5\u30a1\u30ec\u30f3\u30b9\u3067\u306f\u3001Nota AI \u304c\u767a\u8868\u3057\u305f SplitQuant \u306b\u7279\u306b\u611f\u9298\u3092\u53d7\u3051\u307e\u3057\u305f\u3002\u79c1\u305f\u3061\u3068\u3057\u3066\u306f\u3082\u3061\u308d\u3093 MATLAB \u306e <a href=\"https:\/\/www.mathworks.com\/company\/technical-articles\/compressing-neural-networks-using-network-projection.html\">neural network projection<\/a>\u00a0\u3082\u63a8\u3057\u3066\u3044\u307e\u3059\u3002<\/li>\n<\/ul>\n<p>\u30a8\u30c3\u30b8 AI \u30b3\u30df\u30e5\u30cb\u30c6\u30a3\u306e\u7d99\u7d9a\u7684\u306a\u9032\u6b69\u306b\u5fc3\u304b\u3089\u611f\u52d5\u3057\u3066\u3044\u307e\u3059\u3002Edge AI \u30ab\u30f3\u30d5\u30a1\u30ec\u30f3\u30b9\u306e\u3088\u3046\u306a\u30a4\u30d9\u30f3\u30c8\u306f\u3001\u523a\u6fc0\u7684\u3067\u304a\u4e92\u3044\u306e\u5354\u529b\u306e\u7cbe\u795e\u3092\u80b2\u3093\u3067\u3044\u307e\u3059\u3002\u3053\u306e\u5206\u91ce\u306e\u9032\u5c55\u306b\u306f\u671f\u5f85\u304c\u9ad8\u307e\u308a\u307e\u3059\u3057\u3001MathWorks \u304c\u305d\u306e\u767a\u5c55\u306b\u8ca2\u732e\u3057\u3066\u3044\u308b\u3053\u3068\u3092\u8a87\u308a\u306b\u601d\u3044\u307e\u3059\u3002<\/p>\n<p>\u30a8\u30c3\u30b8AI \u5411\u3051\u306e\u63d0\u4f9b\u5185\u5bb9\u306b\u3064\u3044\u3066\u8a73\u3057\u304f\u306f\u3001<a href=\"https:\/\/www.mathworks.com\/solutions\/deep-learning\/embedded-ai.html\">MATLAB and Simulink for Embedded AI<\/a> \u3092\u3054\u89a7\u304f\u3060\u3055\u3044\u3002<\/p>\n<h6><\/h6>\n","protected":false},"excerpt":{"rendered":"<div class=\"overview-image\"><img decoding=\"async\"  class=\"img-responsive\" src=\"https:\/\/blogs.mathworks.com\/deep-learning\/files\/2025\/03\/figure1.jpg\" onError=\"this.style.display ='none';\" \/><\/div>\n<p>\u203b\u3053\u306e\u6295\u7a3f\u306f 2025 \u5e74 3 \u6708 24 \u65e5\u306b Artificial Intelligence \u3078 \u6295\u7a3f\u3055\u308c\u305f\u3082\u306e\u306e\u6284\u8a33\u3067\u3059\u3002<\/p>\n<p>The following blog post is from Jack Ferrari, Edge AI Product Manager, Joe Sanford, Senior Application Engineer, and Reed Axman,&#8230; <a class=\"read-more\" href=\"https:\/\/blogs.mathworks.com\/japan-community\/2025\/04\/11\/edge-ai-foundations-austin-2025-jp\/\">read more >><\/a><\/p>\n","protected":false},"author":159,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[182,76],"tags":[],"_links":{"self":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/13161"}],"collection":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/users\/159"}],"replies":[{"embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/comments?post=13161"}],"version-history":[{"count":12,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/13161\/revisions"}],"predecessor-version":[{"id":13208,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/posts\/13161\/revisions\/13208"}],"wp:attachment":[{"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/media?parent=13161"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/categories?post=13161"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blogs.mathworks.com\/japan-community\/wp-json\/wp\/v2\/tags?post=13161"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}