1
课程详述
COURSE SPECIFICATION
以下课程信息可能根据实际课需要或在课程检讨之后产生变动。如对课程有任何疑问,
联系授课教师。
The course information as follows may be subject to change, either during the session because of unforeseen
circumstances, or following review of the course at the end of the session. Queries about the course should be
directed to the course instructor.
1.
课程名称 Course Title
大数据实证研究 Empirical Research in Big Data
2.
授课院系
Originating Department
信息系统与管理工程系 Department of Information Systems & Management Engineering
3.
课程编号
Course Code
MIS 402
4.
课程学分 Credit Value
3
5.
课程类别
Course Type
专业选修课 Major Elective Courses
6.
授课学期
Semester
春季 Spring
7.
授课语言
Teaching Language
中英双语 Bilingual
8.
他授课教师)
Instructor(s), Affiliation&
Contact
For team teaching, please list
all instructors
卢涛,信息系统与管理工程系
Tao Lu, Department of Information Systems & Management Engineering
lut@sustech.edu.cn
9.
验员/、所、联
方式
Tutor/TA(s), Contact
待公布 To be announced
10.
选课人数限额(可不)
Maximum Enrolment
Optional
2
11.
授课方式
Delivery Method
讲授
Lectures
实验/
Lab/Practical
其它(具体注明)
OtherPleasespecify
总学时
Total
学时数
Credit Hours
32
32
0
64
12.
先修课程、其它学习要求
Pre-requisites or Other
Academic Requirements
MIS306 数据挖掘及商务应用
MIS306 Data Mining and Business Applications
13.
后续课程、其它学习规划
Courses for which this course
is a pre-requisite
None
14.
其它要求修读本课程的学系
Cross-listing Dept.
None
教学大纲及教学日历 SYLLABUS
15.
教学目标 Course Objectives
This course is designed to instruct the theoretical foundations as well as applications about deep learning algorithms for
high grade undergraduate students. From the theoretical perspective: the course will deliver: 1) fundamental machine
learning theories related to deep learning algorithms, and 2) underlying theoretical mechanism of deep learning models.
Also, from the application perspective: the course will guide the students to master: 1) How to use and apply deep
learning models, and 2) How to design and develop models in real life settings.
课程目标是教授本科高年级学生深度学习的基础理论和应用方法。理论基础方面,课程会教授一些深度学习模型会涉及到
的基础机器学习原理以及深度学习的理论原理。应用方面,课程会带领学生学会如何使用一些常用的深度学习模型,并且
指导学生在实际场景中应用深度学习技术。
16.
预达学习成果 Learning Outcomes
After learning the course, students are expected to:
1 Understand all the basic fundamental concepts in deep learning algorithms;
2 Mater the underlying theories and corresponding derivations;
3 Understand the underlying mechanisms of several commonly used model specifications;
4 Be able to use and apply these models into real life settings;
希望学生学完课程之后能够:
1 理解所有深度学习算法的底层概念;
2 掌握深度学习算法的理论和推导过程;
3 理解一些常用深度学习模型的作用机制;
4 能够在一些实际场景下使用和应用深度学习算法。
17.
课程内容及教学日历 (如授课语言以英文为主,则课程内容介绍可以用英文;如团队教学或模块教学,教学日历须注明
主讲人)
Course Contents (in Parts/Chapters/Sections/Weeks. Please notify name of instructor for course section(s), if
this is a team teaching or module course.)
3
理论和实验课(共
64
学时)
理论课(
32
学时)
第一章 绪论(
2
学时)
简要介绍机器学习、主要是深度学习理论和应用的发展轨迹。包含但不限于:脑神经模型、神经网
络全拟合理论、非线性激活函数、并行计算等
第二章 模型评估与选择(
2
学时)
介绍机器学习模型的评估方法,教授各种评估指标的理论基础,结合各种评估标准讲解模型选择的
利弊权衡方法。
第三章 神经网络模型(2 学时)
介绍神经网络模型的基本原理和模型推理计算过程。
第四章 反向传播(
2
学时)
介绍神经网络反向传播的计算原理和过程。
第五章 梯度下降(
2
学时)
介绍机器学习,主要是神经网络模型的训练方法。
第六章 过拟合与正则化(
2
学时)
介绍机器学习,主要是神经网络模型的过拟合问题,并介绍一些常用的对抗过拟合的方法。
第七章 计算学习理论(
2
学时)
介绍机器学习计算学习理论,比如
PAC
学习和
VC
第八章 期中复习考试(
2
学时)
对上半学期学习的理论指数进行一个回顾复习,组织期中考试。
第九章 深度学习模型应用(2 学时)
介绍一些目前深度学习正在应用的业界场景
第十章 卷积神经网络(2 学时)
介绍卷积神经网络的模型结构、计算原理和应用场景
第十一章 循环神经网络(2 学时)
介绍循环神经网络的模型结构、计算原理和应用场景
4
第十二章 残差神经网络(
2
学时)
第十三章 嵌入式模型(
2
学时)
介绍嵌入式模型的模型结构、计算原理和应用场景
第十四章 生成模型(
2
学时)
介绍生成模型的模型结构、计算原理和应用场景
第十五章 前沿理论分享(
2
学时)
介绍一些前沿的机器学习理论和模型实
第十六章 项目汇报(
2
学时)
组织同学作项目报告
实验课(32 学时)
第一单元 基于
python
的深度学习算法基础(
6
学时)
带领学生实
python
代码基础:数据处理、基本机器学习算法实现、模型评价方法实现
第二单元 深度学习模型代码实现(
18
学时)
带领学生实现各项深度学习模型代码:
PyTorch
、模型架构、梯度下降算法、卷积神经网络、循环神
经网络、残差神经网络、嵌入式模型、生成模型
第三单元 算法应用项目(
8
学时)
带领学生小组设计、实现、展示一个应用项目,开发一个可用算法
Lecture and Tutorial (64 hours)
Lecture (32 hours)
Chapter 1 Introduction (2 hours)
Briefly introduce machine learning, especially deep learning theory and applications. Include but not
restricted to: brain model, how neural networks can fit any function, non-liner activation function, parallel
computing.
Chapter 2 Model Evaluation and Selection (2 hours)
Introduce the evaluation methods of machine learning, teach common evaluation methods, and combine
evaluation methods to tell how to select machine learning models.
5
Chapter 3 Neural Network Models (2 hours)
Introduce the fundamental theory and model inference process of deep learning.
Chapter 4 Backpropagation (2 hours)
Introduce the backpropagation theory and the calculation process.
Chapter 5 Gradient Descent (2 hours)
Introduce the training theory and process of the machine learning, especially deep learning.
Chapter 6 Overfit and Regularization (2 hours)
Introduce the overfitting problem in deep learning, and common methods to mitigate the overfitting curse.
Chapter 7 Computational Learning Theory (2 hours)
Introduce the computational theory of machine learning, e.g., PAC learning, VC dimension.
Chapter 8 Midterm Recap and Exam (2 hours)
Give a recap to the theoretical part of the first half semester, and organize the mid-term exam.
Chapter 9 Deep Learning Applications (2 hours)
Introduce some industry applications of deep learning
Chapter 10 Model Specification: CNN (2 hours)
Introduce convolutional neural networks’ (CNN) structure, fundamental computation theory, and the
application scenarios.
Chapter 11 Model Specification: RNN (2 hours)
Introduce recurrent neural networks (RNN) structure, fundamental computation theory, and the application
scenarios.
Chapter 12 Model Specification: Resnet (2 hours)
Introduce residual neural networks’ (Resnet) structure, fundamental computation theory, and the application
scenarios.
Chapter 13 Model Specification: Embedding (2 hours)
Introduce embedding models’ structure, fundamental computation theory, and the application scenarios.
Chapter 14 Model Specification: GAN (2 hours)
Introduce generative adversarial networks’ (GAN) structure, fundamental computation theory, and the
application scenarios.
Chapter 15 Cutting Edge Progresses (2 hours)
6
Introduce the cutting-edge deep learning method and practice
Chapter 16 Projects Report (2 hours)
Organize the students to give project report presentations.
Tutorial (32 hours)
Module 1Python Based Deep Learning Foundations (6 hours)
Lead the students to learn python programming basics: data processing, basic machine learning realization,
model evaluation realization.
Module 2Deep Learning Models Implementations (18 hours)
Lead the students to learn deep learning programming implementations: PyTorch, model structure, gradient
decent, CNN, RNN, Resnet, Embedding models, GAN, and etc.
Module 3Algorithm Application Project (8 hours)
Lead the students to design, realize, and present an algorithm.
18.
教材及其它参考资料 Textbook and Supplementary Readings
Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT press.
Zhang, A., Lipton, Z. C., Li, M., & Smola, A. J. (2021). Dive into deep learning. arXiv preprint
arXiv:2106.11342.
课程评 ASSESSMENT
19.
评估形式
Type of
Assessment
评估时间
Time
占考试总成绩百分比
% of final
score
违纪处罚
Penalty
备注
Notes
出勤 Attendance
10
课堂表现
Participation
10
小测验
Quiz
课程项目
20
平时作业
Graded Team
Assignments
10
期中考试
Mid-Term Quiz
30
期末考试
Final Exam
0
7
期末报告
Final
Presentation
20
其它(可根据需要
改写以上评估方
式)
Others (The
above may be
modified as
necessary)
20.
记分方 GRADING SYSTEM
A. 十三级等级制 Letter Grading
B. 二级记分制(通过/不通过) Pass/Fail Grading
课程审 REVIEW AND APPROVAL
21.
本课程设置已经过以下责任人/委员会审议通过
This Course has been approved by the following person or committee of authority