Course Information
Short Intro
The objective of this course is to provide a complete introduction to
machine learning, which discusses computer programs that automatically
improve their performance through experience (e.g., programs that learn
to recognize faces, recommend music and movies, and drive autonomous
robots).
Description
This course covers theoretical and practical algorithms for machine
learning from different perspectives: theoretical concepts include
induction bias, PAC learning framework, Bayesian learning methods, and
Occam’s razor; programming assignments include hands-on experiments with
various learning algorithms. This course is designed to provide a solid
foundation for students who wish to master modern artificial
intelligence techniques and to provide the necessary methods,
techniques, mathematics, and algorithms for those who wish to pursue
future research in machine learning.
The course will touch on the following topics:
- decision tree learning, linear models
- model selection and optimization, statistical learning methods
- neural networks, deep learning
- nonparametric models, ensemble learning
Keywords: machine learning, statistical learning,
deep learning
Course group
Previous offers: [2024 Autumn], [2024 Spring], [2023
Spring]
Closely related: [Artificial
Intelligence], [Deep Learning], [Natural Language Processing]
Prerequisites
We expect you to have the following skills before taking this
course:
Required:
- Linear algebra (vector, matrix computation, Euclidean spaces).
- Differential calculus (Jacobian, Hessian, chain rule).
- Probabilities and statistics (common distributions, law of large
numbers, conditional probabilities, Bayes)
- Proficient in one programming language (Python recommended).
Recommended:
- Numerical optimization (notion of minima, gradient descent).
- Algorithm analysis (computational costs).
- Specialized knowledge (visual computing, robotics, speech and
language processing).
Teaching Plan
This course is organized into a 12-week session (4 hours per week).
The main contents are listed below:
- Introduction 导言,Preliminaries 预备知识
- Decision Trees 决策树
- Linear Models 线性模型
- Model Selection 模型选择
- (*)Nonparametric Models 非参数化模型
- (*)Ensemble Learning 集成学习
- (*)Statistical Learning Concepts 统计学习概念
- (*)Probability Models: Complete Data 概率模型:完备数据
- (*)Probability Models: Hidden Variables 概率模型:隐变量
- Feedforward Neural Networks 前馈神经网络,Back Propagation
反向传播
- Deep Learning Computation 深度学习计算
- Convolutional Neural Networks 卷积神经网络
- Modern Convolutional Neural Networks 现代卷积神经网络
- Recurrent Neural Networks 循环神经网络
- Modern Recurrent Neural Networks 现代循环神经网络
Tutorials are designed to consistent with lectures:
- Getting Started 实验配置及基础操作
- Decision Trees 决策树
- Linear Regression 线性回归
- Linear Classification 线性分类
- (*)Nonparametric Models 非参数化模型
- (*)Ensemble Learning 集成学习
- Feedforward Neural Networks 前馈神经网络
- Deep Learning Computation 深度学习计算
- Convolutional Neural Networks 卷积神经网络
- Recurrent Neural Networks 循环神经网络
Schedule
Week |
Date |
Lecture |
Handouts |
1 |
2025/09/08 |
绪论:机器学习范式、基本流程 |
[课程信息] [导言] |
|
2025/09/10 |
实验1:安装配置 |
[安装配置] [实验1:基础编程训练] |
2 |
2025/09/15 |
绪论:历史与发展、深度学习 |
[导言] [预备知识] |
|
2025/09/17 |
|
|
3 |
2025/09/22 |
决策树:模型表达力 |
[决策树] |
|
2025/09/24 |
实验2:决策树模型 |
[实验2:决策树] [数据工具] |
4 |
2025/09/29 |
决策树:ID3算法 |
[决策树] |
|
2025/ |
|
|
5 |
2025/ |
线性回归:最小二乘法 |
[线性回归] |
6 |
2025/ |
实验3:线性回归 |
[实验3:线性回归] |
|
2025/ |
线性回归:梯度下降法 |
[线性回归] |
7 |
2025/ |
线性分类:感知机 |
[线性分类] |
|
2025/ |
实验4:线性分类 |
[实验4:线性分类] |
8 |
2025/ |
线性分类:逻辑回归 |
[线性分类] |
|
2025/ |
|
|
9 |
2025/ |
前馈神经网络:反向传播算法 |
[前馈神经网络] |
|
2025/ |
实验5:前馈神经网络 |
[实验5:前馈神经网络] |
10 |
2025/ |
前馈神经网络:多层感知机 |
[前馈神经网络] |
|
2025/ |
|
|
11 |
2025/ |
卷积神经网络:卷积运算 |
[卷积神经网络] |
|
2025/ |
实验6:卷积神经网络 |
[实验6:卷积神经网络] |
12 |
2025/ |
卷积神经网络:构建与参数 |
[卷积神经网络] |
|
2025/ |
|
|
Several special topics:
Ref. [2024 Spring], [2023 Spring].
Methodology
Problem-solving oriented, equal emphasis on lecture and practice.
以解决实际问题为导向,教学与实践并重。
Each lecture is roughly organized into 3 progressive units:
- Core Concepts 核心概念: provides elementary knowledge of the
topic
- Advanced Discussion 进阶讨论: provides in-depth understanding,
mathematical formulations
- Practical Skills 实践技巧: provides problem-solving skills through
hands-on programming training
Evaluation
- Attendance & participation: 10%
- Understanding of topics: 20%
- Practical exercises: 20%
- Final exam/project: 50%
Project
Textbook
Not mandatory but recommended:
- Christopher Bishop, Pattern Recognition and Machine
Learning, 2006, aka. PRML
- Bishop, C.M. and Bishop, H. Deep Learning: Foundations and
Concepts. Springer, 2023. aka. “New Bishop”
- Zhang et al., Dive into Deep Learning, 2023,
即《动手学深度学习》
- Hastie, T., Tibshirani, R., & Friedman, J. The Elements of
Statistical Learning: Data Mining, Inference, and Prediction (2nd
ed.). Stanford, 2009. 即《统计学习要素:
机器学习中的数据挖掘、推断和预测》
- S. Theodoridis, Machine Learning: A Bayesian and Optimization
Perspective, 2nd Edition 2019.
即《机器学习:贝叶斯和优化方法》
- A. Gelman et al., Bayesian Data Analysis, 3rd Edition
2013.
- Murphy, Probabilistic Machine Learning: An Introduction &
Advanced Topics, 2022
Resource
- 动手学深度学习
- Machine Learning with PyTorch and Scikit-Learn Book: GitHub, online courses
- PRML Book
- Deep Learning: Foundations and Concepts
- The Elements of Statistical Learning
- Bayesian Data Analysis
- Probabilistic Machine Learning: An Introduction,
Advanced Topics
- Mathematics for Machine Learning
- 微软人工智能教育与学习共建社区(Microsoft AI Education Community,
简称AI-Edu)