CSE 512: Machine Learning Fall 2024

Time: Mon. & Wed. 2pm-3:20pm (Location: Light engineering lab, 102)

Key online resources: Blackboard (most official), Piazza (basically official), this website (least official)

Virtual?: I am not planning to record lectures this semester.

Syllabus: here

TA: tbd

Auditors: Welcome to join and access all resources on this website. (Sign up for piazza)

Please email me if you cannot resolve any technical issues: ysun at cs dot stonybrook dot edu

Overenrollment: In an effort to help gauge the "long term enrollment level" of this course, please take the time to skim through the giant pile of problems, which are indicative of the level of homeworks you will have to do in the course. Please use this to calibrate your workload, as you plan your courses for the semester.

Am I ready to enroll in this course?

Homework 0: Here Please review this to see if you have the necessary preparation for this course. Also, review the giant pile of problems linked above.

**ENROLLMENT NOTE: Please fill out this google form if you are trying to enroll but the course is full.

Learning goals

Gain mastery of key machine learning tools
Be proficient in discussing important concepts such as: generalization, decision theory, data assumptions, data quality, separability, representational learning, model training, etc
Understand not just how to implement fundamental tools, but the mathematical theory behind their construction and verifiability
Be conversant in the fast-moving latest technology

Extra reading references

Main References: These books may be helpful for certain subjects in the course

Richard O. Duda, Peter E. Hart, and David G. Stork, Pattern Classification.
Kevin Murphy, Probabilistic Machine Learning: An Introduction.
Alberto Leon-Garcia . Probability, Statistics, and Random Processes for Electrical Engineering

Extra References: These books are for those who wish to read beyond the material in the course

Roger A. Horn, Charles R. Johnson. Matrix Analysis.
Stephen Boyd, Lieven Vandenberghe. Convex optimization.
Rick Durret. Probability theory and examples.

Exercises

giant pile of problems from past offering, datasets, iPython notebooks

Tentative subjects

Point estimation, statistical decision theory, naive Bayes, nearest neighbor, linear regression, optimization, logistic regression, support vector machines, multiclass classification, decision trees, generalization, bagging and boosting, kmeans, GMM, learning representations, graphical models, hidden Markov models, dual and kernel SVMs.

Tentative workload

5 homeworks (50%), 2 midterms (50%)