EE 769 – INTRODUCTION TO MACHINE LEARNING

EE 769 – Introduction to Machine Learning

Semester : Spring 2018

Instructor : Prof Amit Sethi

Motivation :

The course aims to introduce the basics of the field of Machine Learning, focus on the core theory aspects along with some practical coding experience.

Course Content :

  • PAC Learnability and Loss
  • VC Dimension
  • Linear Classifiers
  • Logistic Regression
  • SVM and Kernels
  • Decision Trees
  • Neural Networks
  • Clustering
  • Principal Component Analysis
  • Expectation Minimization Algorithm

Prerequisites :

The course assumes familiarity with basic probability and statistics. For the assignments, Python was expected at a certain level, and those students who weren’t experienced in Python struggled.

Feedback on lectures :

Initial lectures were on the board, but due to the large size of the class, Prof. shifted to slides. The slides were self sufficient in the way that nothing outside of what was mentioned in the slides was expected to know for the exams. The sources of the material in the slides was mentioned and was usually one of the 3 books I’ll mention below. But one is expected to do the hard work of gleaning the required information from the various sections in the books. The Prof. is very concise in the class, although one could wish the laconic slides were more apposite to the books.

Feedback on Exams :

The exams were the best part of the course. They were extremely intuitive, and always had bonus mark questions, so getting a good score in the exams was not a big issue if one was clear with the concepts. No rote was required and the exams could be aced with minimal effort if one is already aware of the concepts.

Course Project and Assignments:

This is probably the most important learning part of the course, where one is expected to work on a real life Machine Learning problem. Most of the projects were based on Deep Learning thanks to the DL boom these days. Sincere efforts in the project would lead to good marks. Assignments were also pretty chill and one could easily score full marks, given he/she starts working early enough on them.

Grading and Difficulty level:

Prof. Sethi was pretty clear on the grading that it would be absolute. 100+ would yield AP, 90+ would yield AA, 80+ would yield AB and so on…

Given the extra bonus marks in the project, Mid-Semester, End-Semester and the assignments, one can easily score 90+ given he/she is sedulous and regular. I would rate the course as Moderately Difficult for those who have never been exposed to ML, but Relatively Easy for the others.

Attendance :

The Prof. mentioned in the first lecture that attendance was not compulsory which led to a sharp drop in attendance. The Prof. appeared pretty indifferent to this drop in attendance though.

Books :

Christopher Bishop – Pattern Recognition and Machine Learning

Shai Shalev – Understanding ML

Pattern Classification – Duda, Hart and Stork

Future direction :

I would say this is one of the most useful courses a student can do, in terms of opening avenues to newer job/internship/research profiles. The next logical continuation could be CS 726 – Advanced Machine Learning. Reviewed by : Avineil Jain (avineil96@gmail.com)