機器學習原理及工業應用
Machine Learning and Industrial Application
| 節 | 週五 |
|---|---|
2 09:00–09:50 | 機器學習原理及工業應用 3 節連堂 |
3 10:10–11:00 | |
4 11:10–12:00 |
* 根據陽明交大上課時間表所列
This course is intended to teach participating students the followings 1. to learn to use Python 3 and libraries (學習使用 Python 3) 2. to acquire fundamental concepts of Machine Learning (獲得機器學習的基本概念) 3. to gain exposure of the real-world industrial (job) projects (接觸現實世界的工業(工作)項目) 4. to create successful Machine Learning applications through exercises (including final project) (練習(包括最終項目)創建成功的機器學習應用程序) 5. to present his/her ML project of chosen subject(s) as a project manager (以項目經理的身份展示他/她所選主題的機器學習項目) 6. to defend the result of his/her chosen ML project when being reviewed (criticized) (在審查(批評)時捍衛他/她選擇的 ML 項目的結果) This course is NOT intended to let participating students study the mathematics of each Machine Learning methods (k-Nearest Neighbor, Neural Network, etc.), but instead understand each methods and be able to use them. 本課程並非旨在讓參與的學生學習每種機器學習的數學(k-Nearest Neighbor、神經網絡等),而是要了解並能夠使用每種方法。 Students' presentation on his/her final project to the entire class will constitute a part of his/her course grade. As in the real-world industrial environment, communication skills play a vital role in his/her success of professional career(s) . 學生向全班展示他/她的期末項目將構成他/她的課程成績的一部分。 與現實世界的工業環境一樣,溝通技巧在他/她的職業生涯成功中起著至關重要的作用。 I expect the students can readily join a Machine Learning project in the industry and work on the projects. Some individual student(s) can perhaps conduct his/her own Machine Learning applications (industrial or otherwise). 我預期學生們可以很容易地加入工業中的機器學習項目並從事這些項目。 一些個別學生也許可以進行他/她自己的機器學習應用程序(工業或其他) CONTENT OUTLINE (Tentative, may change) (2022 Fall semester) Machine Learning in the Industrial World. Chapter 4-7 (程式設計與數據分析): Python. Chapter 1: Introduction to Machine Learning Chapter 13-14 (程式設計與數據分析): 數據分析+機器學習 Chapter 2: Supervised Learning. Chapter 3: Unsupervised Learning. Chapter 4: Representing Data and Engineering Features*. Chapter 5: Model Evaluation and Improvement**. (*) taught with Unsupervised Learning (**) taught with Supervised Learning
1. Basic programming experience (Python, Matlab, C, C++, etc.) is recommended. 2. English - must be able to read and write (e.g. PPT), and preferably can carry out some conversation. 3. Math - Basic Algebra or Calculus
Syllabus https://drive.google.com/file/d/1k3iuUw8Mn3C7Yva-Hh3s1b0DmfQEBmW8/view?usp=sharing TA - not decided yet Computer - Students are strongly recommended to have his/her own Windows 10 laptop (PC). Python will be installed. Without personal laptop PC, students are to have access to Windows 10 PC.
Attendance + Quiz (in class) 10% Homework 25% Midterm test or project 25% Final Project & presentation 40%
| 週次 | 主題 |
|---|---|
| 第 1 週 | 1. overall review with syllabus 2. introduction of the class |
| 第 2 週 | 1. continue introduction to the class 2. install Python & set up environment 3. teach Python (data type, print) 4. hand-on practice |
| 第 3 週 | 1. teach Python 2. hand-on practice in class (format print, math/logical/bitwise/function/...) 3. issue HW1 |
| 第 4 週 | 1. teach Python (function, numpy, file IO, plotting, ...) 2. hand-on practice in class |
| 第 5 週 | 1. teach Python 2. due HW1 3. issue HW2 4. run 1st ML example - IRIS 5. underfit & overfit |
| 第 6 週 | 1. Quiz#1 2. Python 3. review/demo Quiz#1 4. Supervised Learning - KNN |
| 第 7 週 | Supervised Learning - KNN algorithm & its parameters; hand-on practice using forge data and breast cancer data. issue HW3. |
| 第 8 週 | review HW3 Supervised Learning - KNN regression hand-on practice using wave data and KNN regression explain cross validation |
| 第 9 週 | midterm linear regression & example Ridge regression (L2 type) & example issue HW4 (must use cross validation) |
| 第 10 週 | Lasso Regression (L1 type) & example ElasticNet Regression (L1 + L2) & example Logistic Regression & example LinearSVC (SVM) & example |
| 第 11 週 | multi-class classification & example Naive Bayes (BernoulliNB) Decision Tree & example Random Forest classification and regression & examples Gradient Boosted Regression & example issue HW5 |
| 第 12 週 | explain 5-folds & review HW5 convey final project and grading method Support Vector Machines - SVC/SVR Neural Networks - MLP industrial application 1 - laser cutting |
| 第 13 週 | explain final project & grading industrial application 2 - tool's wear & tear prediction (correlation), discussion review student's final project topics uncertainty function, probability feature representations unsupervised learning - PCA |
| 第 14 週 | unsupervised learning - NMF, t-SNE unsupervised learning - clustering |
| 第 15 週 | 25 groups of final project presentations |
| 第 16 週 | Final project presentations (rescheduled to 12/23 afternoon) |
| 第 17 週 | flexible arrangement |
| 第 18 週 | flexible arrangement |
1. Introduction to Machine Learning with Python, Andreas C. Muller & Sarah Guido 2. PYTHON 程式設計與數據分析/白文章 編著 (call Mr. Kuo 0921-456015) 3. Hands-on Machine Learning with Scikit-Learn, Keras & TensorFlow, Aurelien Geron 4.Practical Deep Learning 實用深度學習/謝哲光 5. Invent Your Own Computer Games with Python, AI Sweigart
- 地點
- (1) classroom (2) office (Room 444, E5)
- 時間
- (1) Friday after class time, each week (2) by appointment
- 聯絡方式
- e-mail : yearnhwang@nycu.edu.tw yearnhwang@gmail.com office phone : 55246 LINE : intend to set up a line group by TA