### CS 419 – MACHINE LEARNING

**Course:** CS 419 (Minor)

**Semester:** Spring 2018

**Instructor:** Prof. Preethi Jyothi

**Prerequisites:** EE 223 (A basic probability course)

**Course content:**

1.Basic classification and regression techniques – naive Bayes’, decision trees, SVM, bagging & boosting, linear & logistic regression, perceptrons
2.Maximum likelihood estimates
3.Regularization
4.Basics of statistical learning theory
5.Basics of deep learning
6.Dimensionality reduction techniques – PCA & LDA
7.Unsupervised learning – k means clustering, gaussian mixture models
8.Selected topics from natural language processing and computer vision
**Course Webpage:**

https://www.cse.iitb.ac.in/~pjyothi/cs419

**Feedback on lectures:**

The professor teaches well and primarily uses the white board to teach. She encouraged students to attend the lectures and therefore, very brief slides were provided. The pace of lectures is moderate but gets faster at the end of the course. The course is more inclined towards theoretical side of machine learning and requires one to be familiar with concepts from probability and linear algebra. However, the professor did cover basics of applications like computer vision and speech recognition.

There were two interesting guest lectures as well –

1.Prof. Amit Sethi on “some cases of pathology diagnostics using ML” 2.Prof. Arjun Jain on “Deep Learning for Computer Vision” Feedback on assignments:

There were total four assignments for the course. The first two had both theoretical and programming based questions. The third one was entirely theoretical and the fourth one was totally programming based. The assignments were a little difficult. It is good to start early and attempt both the programming based and theoretical questions. The number of attempts to programming based questions reduced dramatically after the first assignment. Having familiarity with Python and Numpy will be beneficial.

**Feedback on project:**

Students had to do a course project in groups of 2 to 4. A brief abstract had to be submitted by the end of first month of teaching which was aimed to get students to form groups and finalize the problem statement. A short progress report had to be submitted a month before the endsems. The final presentation and report were due after the endsems. The professor is very considerate and did give credit for failed attempts to solve the problem.

**Feedback on exams:**

The evaluation breakdown is given below –

Assignments – 40% Mid Semester Exam – 15% Project – 15% Class participation (mini quizzes) – 5% End Semester Exam – 25% There were mini class quizzes which contributed to class participation score. The midsem and the endsem exams were on the easier side. Some of the mathematically heavier topics like the EM algorithm were not even touched in the exams. The exams were open notes. Having some familiarity with the content was enough to score well in the exams.

**Difficulty:** Moderately difficult

**Grading:** Check on asc.

**Study material and references:**

1.Pattern Recognition and Machine Learning. Christopher Bishop
2.Foundations of Data Science. Avrim Blum, John Hopcroft and Ravindran Kannan. January 2017
**Reviewed by** – Anmol Kagrecha, 4th year Dual Degree (CSP)