You are on page 1of 9

E1 277: Reinforcement Learning

Instructor: Gugan Thoppe

Teaching Assistants
Ankur Naskar
Kaustubh Kartikey
Naman
Nibedita Roy
Administrative Matters

1
Microsoft Team Details

• Name: E1 277 Reinforcement Learning (2023-24)

• Team code: 7evboz0

• Discussions, assignments, exam papers, etc. would be on Teams

2
Course Schedule

• First Class: 09th Jan. 2024

• Last Class: 11th Apr. 2024

• Total Lectures: 28

3
Grading Criteria

Assignments (04) 20

Midterms (02) 30

Project (01) 20

Final exam (01) 30

4
Important Dates

# Release Due
A1 26 Jan 09 Feb

A2 09 Feb 23 Feb

M1 16 Feb 16 Feb

A3 08 Mar 22 Mar

A4 22 Mar 05 Apr

M2 29 Mar 29 Mar

P 12-13 Apr 12-13 Apr

F TBD TBD

5
Syllabus

• Introduction to RL

• Preliminaries: Markov chains, Markov Decision Processes (MDPs)

• MDP formulations: Stochastic shortest path, discounted, average

• RL analysis: Stochastic approximation, Martingale

• RL algorithms: tabular, function approximation, policy gradient,


natural policy gradient

Advanced (if time permits): Deep RL, actor-critic, TRPO, PPO

6
Textbooks (other than my lecture notes)

• Markov Chains by Norris

• Neuro-Dynamic Programming by Bertsekas and Tsitsiklis

• Dynamic Programming and Optimal Control-I, II by Bertsekas

• Algorithms for Reinforcement Learning by Szepesvari

• Reinforcement Learning: An Introduction by Sutton and Barto

You might also like