Course Description

Machine learning is at the core of the emerging "Data Science", a new science area that promises to improve our understanding of the world by analysis of large-scale data in the coming years. The objective of this class is to provide a rigorous training on the fundamental concepts, algorithms, and theories in machine learning. The lectures will cover perceptrons/linear models, projection/nonlinear embedding methods, neural networks/deep learning, parametric/non-parametric methods, kernel machines, mixture models and graphical models. Various examples and applications will also be discussed in demos. The specific goals for students include

• To learn key concepts in machine learning.
• To acquire hands-on experiences with implementation of machine learning algorithms.
• To understand how to derive the mathematical formulation of the fundamental machine learning models.
• To learn how to formulate and solve application questions with appropriate machine learning methods.

Course Design

This course will be primarily lecture-based. The breakdown of the class grade is as follows:

• Homework assignments (50%): There will be six hands-on homework assignments (no grade for homework 0). Each requires substantial work in programming and math derivation. The programming assignments must be completed in Python. Each homework will be due at 11:59 PM CDT. For late submissions, each student is allowed to submit a maximum of two out of the six assignments late with a maximum delay of three days for each late submission. Any submissions that do not follow this policy will not be graded and will receive zero credit -- this rule will be strictly enforced. If there are questions, contact the instructor and the TA.
• Midterm exam (25%): 1-hour in-class closed-book exam.
• Final exam (25%): 1-hour in-class closed-book exam.
• Due to the in-class exams, we will NOT be able to offer alternative exam time other than accommodating the need of disability services. Thus, we will NOT allow class time conflict.

Textbooks

Introduction to Machine Learning (Third Edition) (Links to an external site.)Links to an external site., Ethem Alpaydin, MIT Press, ISBN: 9780262028189

Pattern Recognition and Machine Learning (Links to an external site.)Links to an external site., Christopher Bishop, Springer, ISBN: 978-0-387-31073-2 (Secondary)

Prerequisite

Programming with Matlab, and prior knowledge in basic statistics, probabilities and linear algebra. 

The machine learning class is built on prior knowledge in programming, algorithms, statistics/probability theory and linear algebra. Here are some good resources for your preparation or assessment of your readiness for the class.

• Matrix calculus and linear algebra
• Probability theory
• Python programming tutorial

After going over the above materials, you can assess your readiness for the class in homework 0.

If you found the problems very beyond your understanding, you might have difficulty in passing the class.

Time and Location

Tuesday/Thursday, 1:00pm-2:15pm, Peik Hall 28

Academic Integrity Policy

Students are encouraged to discuss the homework assignments with each other, but each student must complete and submit his/her own work. Any student cheating on a homework assignment will receive an F as a class grade and the incident will be reported to the University office. Group work will be encouraged on the course project. More information on academic misconduct is available at Note on Academic Conduct for New Students and The Office for Student Academic Integrity.