本課程目標在於訓練同學快速開發物聯網智慧應用,以聯邦學習進行閘道器分散式訓練,配合kubeflow管線式開發,練習建置聯邦學習容器,導入kubeflow自動部署,進行聯邦模型參數自動更新(CI/CD),最後結合Node red 圖形式開發kubeflow 聯邦學習管線,供工程師快速開發物聯網智慧應用,並以模板技術(template)結合容器式開發,實作各種聯邦學習,包含縱向、橫向及聯邦遷移,最後貢獻於Node-red及kubeflow開源社群。The purpose of this course is to train students to quickly develop smart applications of the Internet of Things, conduct gate dispersed training with federal learning, cooperate with kubeflow pipeline development, practice building of federal learning containers, introduce kubeflow automatic deployment, conduct automatic update of federal model parameters (CI/CD), and finally develop kubeflow in Node red figure form The federal learning pipeline is designed for engineers to quickly develop smart applications of the Internet of Things, and develop in container-based form with template technology, implementing various federal learning, including directional, directional and federal migration, and finally paid for the Node-red and kubeflow origin community.
1. Maneesh Rao, Internet of Things with Raspberry Pi 3, Apr. 2018.
2. 本計畫課程模組B-1,家電與能源管理應用開發模組教材
3. 本計畫課程模組B-4 ,深度學習時空間資料探勘教材
4. 本計畫課程模組B-3,家電與能源管理應用開發模組教材
5. 其他參考論文或網路上資料之自編教材
1. Maneesh Rao, Internet of Things with Raspberry Pi 3, Apr. 2018.
2. This plan course module B-1, home appliance and energy management application development module textbook
3. This planning course module B-4, in-depth learning of time and space data exploration textbook
4. This plan course module B-3, home appliance and energy management application development module textbook
5. Self-edited textbooks for other reference papers or online data
評分項目 Grading Method | 配分比例 Grading percentage | 說明 Description |
---|---|---|
quiz1quiz1 quiz1 |
15 | home appliance& power management module |
midtermmidterm midterm |
20 | deep learning module |
quiz2quiz2 quiz2 |
15 | Social media and line bot |
term projectterm project term project |
30 | social bot IOT |
HomeworkHomework homework |
20 | Rasberry PI homework |