EE 559 syllabus Page 1 v1 EE 559 Machine Learning I: Supervised Methods Units: 4 Lecture: MW 3:30 – 5:20 PM Pacific Time Discussion: Thu 5:30 – 6:20 PM Pacific Time Location: OHE 122, online, and DEN@Viterbi Course URL: https://courses.uscden.net/d2l/home Instructor: B. Keith Jenkins Office Hours: TBA Contact Info: Email: [email protected] [Please include “EE 559” in the subject line] Online: piazza.com Teaching Assistants: TBA IT Help: For help with coding machine-learning algorithms, consult piazza, online forums, or the TAs; for help with other python coding, working with datasets, or using library routines in python, also try online help and documentation, and reference resources given below; for help with USC-supplied software or on-campus networking, consult USC ITS at https://itservices.usc.edu/contact/ .EE 559 syllabus Page 2 v1 Course Description (Catalogue) Distribution-free and probabilistic methods for supervised classification and regression; learning algorithms; optimization techniques; feature-space transformations; parametric and nonparametric methods; Bayes decision theory; artificial neural networks. Course Description (Expanded) Concepts and algorithms for pattern recognition and regression using machine learning are covered in depth. The course will stress an understanding of different supervised-learning algorithms at both theoretical and practical levels, as well as their advantages and disadvantages. Underlying fundamentals are emphasized, including theory and origins of learning algorithms and criterion functions. The goal is to give the student an understanding of some fundamental approaches to machine learning, to enable further study and growth on their own. The student’s work will include mathematical analysis, analytical understanding, and writing and running code that learns from data. A moderately sized project in the second half of the semester will involve developing and optimizing one or more machine learning systems to perform well on real-world datasets. This course is intended for graduate students in Electrical and Computer Engineering or related fields, who wish to gain an understanding of, and some experience with, machine learning approaches, tools, and techniques. Only supervised-learning methods are covered in this course. Learning Objectives After successfully completing this course, the student will: • Have a perspective of different approaches to supervised machine learning for pattern classification and regression • Understand the underlying math of a variety of supervised-learning methods • Be able to use statistical and non-statistical techniques to solve machine learning problems • Be able to code and run algorithms for learning from data • Be able to optimize machine learning algorithms and systems, and assess their overall performance • Know how to create new techniques for machine learning where needed Prerequisite(s): Co-Requisite(s): EE 503 and EE 510 Concurrent Enrollment: Recommended Preparation: knowledge of python (e.g., at the level of EE 541), knowledge of multivariate calculus (at the sophomore or junior undergraduate level). For students that don’t know python, resources will be provided to help them learn python during the first few weeks of the semester, leveraging from their knowledge of MATLAB; students should allow some extra time for this if they don’t know much python. Course Notes The Desire2Learn (D2L) system will host the course website. On the website will be posted video recordings of the lectures and discussion sessions, lecture and discussion-session notes, additional instructor-provided notes and handouts, assignments and solutions. Your graded assignments and scores will also be visible to you on the website. Technological Proficiency and Hardware/Software Required Python will be used throughout this class for homework assignments and the class project. All students will be responsible for installing and maintaining their own python distribution.EE 559 syllabus Page 3 v1 Required Readings and Supplementary Materials Required text • C. M. Bishop, Pattern Recognition and Machine Learning (Springer, 2006). ISBN-13: 978-0387-31073-2. Available from USC bookstore, Amazon.com, and Springer.com. Supplementary texts • Kevin Murphy, Probabilistic Machine Learning: An Introduction (MIT Press, 2022). Preprint available for download at https://probml.github.io/pml-book/book1.html . • R. O. Duda, P. E. Hart, and D. G. Stork, Pattern Classification, Second Edition (Wiley-Interscience, John Wiley and Sons, Inc., New York, 2001) • T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Second Edition (Springer, 2009) • Simon Haykin, Neural Networks and Learning Machines, 3rd Edition, (Pearson, 2009) • Ethem Alpaydin, Introduction to Machine Learning, Fourth Edition (MIT Press, 2020) • I. Goodfellow, Y. Bengio, and A. Clourville, Deep Learning (MIT Press, 2016) Supplementary resources for Python • Hans Fanghor, Introduction to Python for Computational Science and Engineering (2016) [PCSE], available for free download at: https://github.com/fangohr/introduction-to-python-for-computational-science-and-engineering • Fabio Nelli, Python Data Analytics, (APress, 2015) [PDA], available for download from USC Library: http://usc.summon.serialssolutions.com/search?q=Fabio%20Nelli%2C%20Python%20Data%20Analytics#!/search?ho=t&l=en&q=Fabio%20Nelli,%20Python%20Data%20Analytics • The Python 3 Tutorial [PT]: https://docs.python.org/3.8/tutorial/index.html • EU Python 3 Tutorial [EUP]: Good for Chapters on object oriented programming, class vs. instance attributes, and inheritance http://www.python-course.eu/python3_course.php Description and Assessment of Assignments 1. Homework assignments Homework assignments will, on average, consist of approximately 50% computer problems and 50% analytical problems. Homework assignments will be posted on the course website when assigned, and typically due 1 week later (due date and time will be specified on the homework assignment). Each homework assignment will be worth 10 points. Starting with Homework 3, a portion of the homework score