4190.666 Fall 2017
- Instructor: Hyun Oh Song [email]
- Location: 302-106
- Time: 3:30 - 4:45 Mon, Wed
- Grading TAs: Jaemin Cho, Soochan Lee
- Email: email@example.com
- Office hour: 302-107 (next room), 5:00 - 6:00 Mon
- Auxiliary textbook:
(1) Machine Learning A Probabilistic Perspective-Kevin Murphy-MIT press-2012
(2) Pattern Recognition and Machine Learning-Christopher Bishop-Springer-2006
- Homework assignments (40%), Midterm exam (30%), Final exam or project (30%).
- Final grades will be assigned based on the earned scores.
A: 3-40%, B: 4-50%, C or below: 20% (subject to change according to the University's rule).
- Assignments will be posted via ETL.
- If you have any homework questions, please use the board at ETL.
- Submit pdf file to ETL. No paper submission accepted.
- No late homework accepted. We’ll drop your lowest homework grade instead.
final_hw_score = (\sum_i score(hw_i) - \min_i score(hw_i)) / (num_hw - 1)
- Students are allowed to discuss with others about the problems, but must hand in their own answers.
- This course provides a comprehensive introduction to machine learning. Topics will include supervised learning, unsupervised learning, learning theory, dimensionality reduction, reinforcement learning. Throughout the course, we will draw connections to application areas such as robotics, computer vision, and language processing. If time permits, we will cover extensions to deep learning.
- Students are expected to be familiar in linear algebra, multivariate calculus, and basic probability and statistics. Also familiarity with Matlab, Numpy or a related matrix-oriented programming language is strongly recommended.
- Most of the class material is based on Dr. Alex Smola's machine learning class that the instructor took as a graduate student at UC Berkeley.
- Part of the class material is from Prof. Kyunghyun Cho's machine learning class, Prof. Peter Bartlett's CS281 class at UC Berkeley, Prof. David Silver's reinforcement learning class.
- We greatly thank Prof. Alex Smola, Prof. Kyunghyun Cho, Prof. David Silver, and Prof. Peter Bartlett for generously allowing us to use/modify the course material.