課程名稱 |
統計學習與深度學習 Statistical Learning and Deep Learning |
開課學期 |
109-1 |
授課對象 |
學程 商業資料分析學分學程 |
授課教師 |
盧信銘 |
課號 |
IM5056 |
課程識別碼 |
725 U3670 |
班次 |
|
學分 |
3.0 |
全/半年 |
半年 |
必/選修 |
選修 |
上課時間 |
星期四2,3,4(9:10~12:10) |
上課地點 |
管二305 |
備註 |
商業資料分析學分學程課程 總人數上限:70人 |
Ceiba 課程網頁 |
http://ceiba.ntu.edu.tw/1091sldl |
課程簡介影片 |
|
核心能力關聯 |
核心能力與課程規劃關聯圖 |
課程大綱
|
為確保您我的權利,請尊重智慧財產權及不得非法影印
|
課程概述 |
Statistical learning refers to a set of tools for modeling and understanding complex datasets. It is a recently developed area in statistics and blends with parallel developments in computer sciences and machine learning. The field encompasses many methods such as the regularized regression, classification, graphic models, and recently deep learning. The goal of this course is to introduce a set of tools for data analytics. We will cover the principles and applications of these models/tools. These tools will not be viewed as black boxes. Instead, students will be exposed to the details, not just the use, of these tools. The main reason is that no single approach will perform well in all possible applications. Without understanding how a tool work, it is impossible to select the best tool. This course is appropriate for master's students and advanced undergraduates who wish to use statistical learning, machine learning, and deep learning to analyze their data. |
課程目標 |
Omitted. |
課程要求 |
*Grade Distribution*
The course grades will be determined by the following percentages:
Homework (Python-based) (5-6 Assignments) 55%
Attendance, participation & quizzes 15%
Final Project (Presentation) 30%
Total 100%
*Computational Tools*
Students are required to use Python 3 (with scikit-learn, pandas, matplotlib, numpy, pytorch, etc.) to solve homework assignments.
*Homework*
There are at least five graded assignments. Unless otherwise stated, students are required to organize their code and results using Jupyter Lab and submit their homework to NTUCOOL using the IPYNB format. An assignment is due at the beginning of the first class in the following week. Late submissions will not be accepted. Homework assignments play a very important role in the learning process, and students are expected to spend a significant amount of time in solving homework problems. Students are allowed to discuss about homework questions. However, each student must turn in her/his own homework. Plagiarism will result in severe penalty for everyone involved.
*Final Project (Team-based)*
Students are expected to form teams of three to six people and work on a data analytics problem that is interesting and challenging for you. Details will be given in class.
|
預期每週課後學習時數 |
|
Office Hours |
|
指定閱讀 |
* Pattern Recognition and Machine Learning by Christopher M. Bishop; ISBN 0-387-31073-8.
* Hands-on Machine Learning with Scikit-Learn & Tensorflow by Aurelien Geron; ISBN 978-1-491-96229-9.
* Deep Learning by Ian Goodfellow, Yoshua Bengio, and Aaron Courville; https://www.deeplearningbook.org/
* Dive into Deep Learning by Aston Zhang, Zack C. Lipton, Mu Li, Alex J. Smola; https://d2l.ai/ and https://github.com/dsgiitr/d2l-pytorch
|
參考書目 |
待補 |
評量方式 (僅供參考) |
|
週次 |
日期 |
單元主題 |
第1週 |
9/17 |
Introduction, Regression Part 1 (K-nearest-neighbors) |
第2週 |
9/24 |
Regression Parts 2-3 (Linear models and regularization) |
第3週 |
10/01 |
Holiday, no class |
第4週 |
10/08 |
Regression Part2 4-6: Dummy coding, bias-variance trade off, and an example |
第5週 |
10/15 |
Linear Models for Classification |
第6週 |
10/22 |
Model Evaluation |
第7週 |
10/29 |
Feature Selection |
第8週 |
11/05 |
Dimension Reduction |
第9週 |
11/12 |
Tree-based Models, Part 1 |
第10週 |
11/19 |
Tree-based Models, Part 2: Bagging, Random Forest, Boosting, and Stacking |
第11週 |
11/26 |
Deep Feedforward Networks |
第12週 |
12/03 |
Distributed Representations for Natural Languages |
第13週 |
12/10 |
Regularizations and Optimizations |
第14週 |
12/17 |
Convolutional Network, Part 1 |
第15週 |
12/24 |
Convolutional Network, Part 2 |
第16週 |
12/31 |
Final Project Presentation |
第17週 |
1/07 |
TBD |
|