CS 419 Introduction to Machine Learning ( 2018-2019 )

Instructor: Sunita Sarawagi

Course Name: Introduction to Machine Learning( CS419 )

Course Type: Theory ( Minor )

Credits: 6

Prerequisites: None

Course Content: 

Decision tree classification(Pruning and Regression Trees), Probabilistic classifiers(LDA, QDA, Naive Bayes, Logistic), Hyperplane classifiers(Loss functions), Convex Optimization, Feedforward Neural networks, Convolution Neural Networks, Recurrent Neural Networks, Clustering(k-means, EM algorithm), Combining models(Bagging and boosting), Support vector machines, Overview of Markov Decision Process and Reinforcement Learning. 

Books/Other Resources:

Understanding Machine Learning. Shai Shalev-Shwartz and Shai Ben-David. Cambridge University Press. 2017, Pattern recognition and machine learning by Christopher Bishop, Springer Verlag, 2006, T. Mitchell. Machine Learning. McGraw-Hill

About Lectures:

There was no attendance policy however there were 3 surprise quizzes (best 2 out of 3 ). Initially, the prof used slides to teach, later tablet based teaching and finally she used the blackboard too. There were occasional demos in class. The lectures had some amount of math in them. But it wasn't very hard to follow. Prof gave a 10 min break in the middle of the 90 min lecture which was always a welcome break.

Course Structure:

There were 3 coding assignments. The last assignment was optional where one could do a project (on a self-chosen topic) and present it later. The prof provided the rough template for all three assignments. There was a Kaggle competition hosted for the first 2 assignments, which encouraged students to compete with each other for accuracy. Difficulty (in assignment) : Moderate(for new coders), Easy(others).

25% Mid-semester exam, 40% End semester exam, 20% homeworks, 15% In class quizzes(best 2 out of 3)

Pro tips:

Code as much as you can, going to the lectures is important as the intricate points are discussed there which will help in the assignments as well. The exams are based on fundamentals, just do your theory properly, it will be enough. The prof gave practice problem sets as well, that help.

Advanced follow-up courses ( Courses that can be taken after this ):
ML + IP, data mining, etc.

Respondent: Ananay Garg