CS420 Machine Learning 2018 

Weinan Zhang, Assistant Professor

Zhiyuan College
Shanghai Jiao Tong University

Email: wnzhang [AT] sjtu.edu.cn

Machine learning is the science of training machines with non-explicit programming based on a dataset to get them work on intelligent tasks. Machine learning has obtained fast development during the last two decades and now plays an important role in various aspects of our daily life, such as weather forecasting, e-commerce personalized recommendation, news categorization, face recognition, speech QA, self-driving, home robots, and medical expert system etc. Particularly, in March 2016, Google's AlphaGo beat Lee Se-dol on Game of Go with the score 4-1, which indicates the arrival of the new artificial (general) intelligence era with machine learning based on big data.

This course provides a comprehensive introduction of the fundamental problems and methodologies of machine learning, including supervised/unsupervised learning (covering most prediction applications, e.g., recommender systems, image recognition and webpage ranking etc.) and reinforcement learning (covering all decision-making applications, e.g., playing Go, self-driving, ad bidding and smart stock picking etc.). Additionally, the coursework includes hands-on tasks, in which the students are required to design machine learning programs to accomplish several intelligence tasks, and are highly encouraged to further improve the machine performance via trying different models and refining the code implementation.

Notice

Zhanghao Wu, Yikai Li and Zhou Fan won the outstanding poster award in the poster session on Jun. 11, 2018.

Course Works


link
Course work 1: Text Classification
To create a classification model for recommending selected articles.
Mar. 13, 2018 - Apr. 22, 2018.

link
Course work 2: Item Recommendation
To create a collaborative filtering based recommender system to predict user-item ratings.
Apr. 16, 2018 - May. 21, 2018.

link
Course work 3: Multi-Agent Battle Game AI
To create artificial collective intelligence for multi-agent battle game.
Apr. 30, 2018 - Jun. 11, 2018.

Slides


pdf
Lecture 1: Introduction to Machine Learning
Introduction about AI, machine learning, data science and various ML applications
Feb. 26, 2018

pdf
Lecture 2: Linear Models for Supervised Learning
Discriminative/generative models, linear regression, logistic regression
Mar. 5, 2018

pdf
Lecture 3: Support Vector Machine and Kernel Methods
Maximizing Margin Classification, SVMs, Kernel Tricks
Mar. 12, 2018

pdf
Lecture 4: Neural Networks
Perceptron, Multilayer Perceptron, Backprop, Deep Learning
Mar. 19, 2018

pdf
Lecture 5: Tree Models
Decision Trees, ID3, CART
Mar. 27, 2018

pdf
Lecture 6: Ensemble and Boosting Algorithms
Ensemble methods, Bagging, Boosting
Apr. 2, 2018

pdf
Lecture 7: Ranking and Filtering
Learning to rank, collaborative filtering, matrix factorization
Apr. 9, 2018

pdf
Lecture 8: Probabilistic Graphical Models
Bayes Network, Markov Network, Conditional Independence, Message Passing
Apr. 16, 2018

pdf
Lecture 9: Unsupervised Learning
PCA, Mixture Gaussians, EM Methods, Auto-encoders, GANs
Apr. 23, 2018

pdf
Lecture 10: Learning Theory and Model Selection
PAC Learning Theory, VC Demension, Bias Variance Decomposition, Feature Selection
Apr. 28, 2018

pdf
Lecture 11: Introduction to Reinforcement Learning
Reinforcement Learning, Markov Decision Process, Dynamic Programming, Model-free RL
May. 14, 2018

pdf
Lecture 12: Approximation Methods in RL
Value Approximation Methods, Policy Gradients, Deep RL
May. 21, 2018

pdf
Lecture 13: Transfer Learning
Transfer Learning, Domain Adaptation, Importance Sampling
Jun. 4, 2018

pdf
Guest Lecture: ACTRCE: Augmenting Experience via Teachers' Advice
by Yuhuai Wu from University of Toronto
May 28, 2018

pdf
Guest Lecture: Deep Reinforcement Learning for Robotics: Frontiers and Beyond
by Shixiang Gu from University of Cambridge
May 28, 2018

pdf
Guest Lecture: Communication in Multi-agent Reinforcement Learning
by Ying Wen from University College London
May 28, 2018

pdf
Guest Lecture: Learning and Modeling from Sequential Events
by Prof. Junchi Yan from Shanghai Jiao Tong University
Jun. 4, 2018

Related Readings

Teaching Assistants


link
Jiacheng Yang, ACM15 student, research intern in ApexLab
Research on AutoML, reinforcement learning
Email: kipsora [AT] gmail.com

link
Lianmin Zheng, ACM15 student, research intern in ApexLab
Research on machine learning systems, multi-agent reinforcement learning
Email: mercy_zheng [AT] sjtu.edu.cn

Last Year Course


link
CS420 Machine Learning 2017
Compared to CS420 2017, CS420 2018 adds more materials on (multi-agent) reinforcement learning.

News


Feb. 26, 2018
Course started!