Type of Credit: Elective
Credit(s)
Number Registered
This lecture aims at teaching students the fundamentals and hands-on skills in machine learning and deep learning, in particular supervised learning techniques. Students will be required to use Python exclusively in all the homework assignments.
能力項目說明
The objective of this class includes understanding data representation and underlying mathematical formulation as well as turning mathematical formulation into code and understanding/interpreting the results.
Week 1
Subject：Introduction & syllabus
Covering topics: Introduction to digital image processing.
Reading: N/A
Teaching/HW: Explain the syllabus
Hours spent for preview: N/A
Hours spent for review: 1 hour
--
Week 2 & 3
Subject：Mathematical tools
Covering topics: Linear Algebra and Probability
Reading: Course slides
Teaching/HW: Get familiar with math tools often used in machine learning
Hours spent for preview and review: 2 hours each
--
Week 4, 5, & 6
Subject：Machine Learning Basics
Covering topics: linear and logistic regression/classification, SVM, kNN
Reading: Course slides
Teaching/HW: Introduce various machine learning techniques. HW1 will be released. It needs to be turned in within one week after being released.
Hours spent for preview and review: 2 hours each
--
Week 7
Subject：Optimization and gradient descent
Covering topics: Convex functions and various gradient descent approaches
Reading: Course slides.
Hours spent for preview and review: 2 hours each
Week 8
Subject：Artificial Neural Networks
Covering topics: Perceptron, Multilayer Perceptron, Back-propagation
Reading: Course slides
Teaching/HW: HW2 will be released. It needs to be turned in within one week after being released.
Hours spent for preview and review: 2 hours each
Week 9
Subject: Midterm Exam
Week 10
Subject: Deep learning frameworks
Covering topics: Keras, Tensorflow and Pytorch
Reading: Course slides
Teaching/HW: Get familiar with deep learning frameworks
Hours spent for preview and review: 2 hours each
Week 11 & 12
Subject: Convolutional neural networks
Covering topics: Convolutional neural networks, often-used architectures
Reading: Course slides
Teaching/HW: Final project will be released.
Hours spent for preview and review: 2 hours each
Week 13 & 14
Subject: Recurrent neural networks
Covering topics: Recurrent neural networks (RNNs)
Reading: Course slides
Teaching/HW: Teach student RNNs and their underlying math
Hours spent for preview and review: 2 hours each
Week 15
Subject: Generative adversarial networks
Covering topics: Introduction to generative adversarial networks
Reading: Course slides
Hours spent for preview and review: 2 hours each
Week 16 & 17
Subject: Deep Learning Applications
Covering topics: Computer vision and image processing based on deep learning
Reading: Course slides
Teaching/HW: Discuss several problems better solved by deep learning
Hours spent for preview and review: 2 hours each
Week 18: Final Presentation
1. Homework (30%) - two homework assignments
2. Midterm Exam (30%)
3. Final Project (40%)
Lecture notes and slides
| 書名 Book Title | 作者 Author | 出版年 Publish Year | 出版者 Publisher | ISBN | 館藏來源* | 備註 Note |
|---|
1. Deep Learning http://www.deeplearningbook.org/ 2. Neural Network and Deep Learning http://neuralnetworksanddeeplearning.com/ 3. Deep Learning: A Practitioner’s Approach https://www.amazon.com/dp/1491914254?tag=inspiredalgor20