Review by

Aadesh Anil Madnaik 2022 (BTech.)

Course Offered In Spring 2022 (Jan-Apr 2022)

Instructors Profs D Manjunath & Jayakrishnan Nair

Prerequisites

Basic Probability course like EE325

Difficulty

On a scale of 1 to 5 with 1 being easiest, 1

Course Content

Bandit algorithms, sampling, regret for adversarial, stochastic and Markovian bandits, Thompson Sampling - slides linked [here]{.ul}

Feedback on Lectures

The lectures were quite good overall. Both professors teach well and expect good engagement from the students. The subject taught is fairly easy to understand particularly if you have prior experience with bandit theory from CS747 or any online learning course. The lectures dealt with details of bandit algorithms, formulations of regret, and evaluating policies. It is notable that this course does NOT cover MDPs.

Feedback on Evaluations

Only one mode of evaluation - take-home endsem with 100% weightage

The professors were about to place certain weightage on class interaction, but they removed it after midsem.

Study Material and References

Slides linked [here]{.ul}

Follow-up Courses

Any other courses on online learning, including MDPs will be a good follow-up to this

Final Takeaways

This is a very interesting course and the professors certainly teach it well. The course content is highly important for online learning.

Grading Statistics: