B.E. (Artificial Intelligence and Data Science) Reinforcement Learning Lab Syllabus - Mumbai University
This is the Fourth Year BE AI and DS syllabus under CBCS REV-2019 'C' Scheme, in force from the academic year 2023-24. The University has published no NEP 2020 syllabus for Semesters V to VIII of any engineering branch, so this is the scheme you are examined on — exam form 1T01817 and 1T01818. The first and second years of the degree are on NEP 2020.
Loading syllabus...
Syllabus for Reinforcement Learning Lab
Suggested List of Experiments
- 1 Implementing a simple grid-world environment and training an agent using basic Q-learning
- 2 Implementing a multi-armed bandit problem and comparing different exploration strategies like epsilon-greedy and UCB. 3, Implementing a basic grid-world environment as an MDP and applying policy iteration and value iteration algorithms to find optimal policies.
- 4 Applying dynamic programming algorithms, such as policy evaluation and policy improvement, to solve a small-scale MDP problem.
- 5 Implementing Monte Carlo control and Temporal Difference (TD) learning algorithms to train an agent in a grid-world environment.
- 6 Exploration vs. Exploitation Trade-off: Experimenting with different exploration strategies and analyzing their impact on the learning performance of an agent in a bandit problem.
- 7 Function Approximation in Reinforcement Learning: Using function approximation techniques, such as linear regression or neural networks, to approximate value functions in reinforcement learning problems.
- 8 Deep Reinforcement Learning: Implementing a deep Q-network (DQN) to train an agent to play a popular Atari game, such as Pong or Space Invaders.
- 9 Transfer Learning and Multi-Task Reinforcement Learning: Investigating transfer learning in reinforcement learning by training an agent in one environment and transferring its knowledge to a different but related environment
- 10 Policy Gradient Methods: Implementing policy gradient methods, such as REINFORCE or Proximal Policy Optimization (PPO), to train an agent in a continuous control environment. *11. Applications and Case Studies: Applying reinforcement learning techniques to solve a real-world problem, such as training a self-driving car to navigate a simulated road environment.
Text Books
- 1 Reinforcement Learning: An Introduction, by Richard S. Sutton and Andrew G. Barto
- 2 Alessandro Palmas, Dr. Alexandra Galina Petre, Emanuele Ghelfi, The Reinforcement Learning Workshop: Learn how to Apply Cutting-edge Reinforcement Learning Algorithms to a Wide Range of Control Problems, 2020 Packt publishing.
- 3 Phil Winder, Reinforcement Learning Industrial Applications with Intelligent Agents, O'Reilly
- 4 Dr Engr S M Farrukh Akhtar, Practical Reinforcement Learning, Packt Publishing, 2017.
Useful Links
- 1 Machine Learning and Friends at Carnegie Mellon University
- 2 Reinforcement Learning: A Survey
- 3 Bibliography on Reinforcement Learning
- 4 David J. Finton's Reinforcement Learning Page
Reproduced from the University of Mumbai syllabus for B.E. (Artificial Intelligence and Data Science), item 6.12 (N), under CBCS REV-2019 'C' Scheme, in force from the academic year 2023-24. Wording, module numbering and hours are as printed in that syllabus.
The complete syllabus
This subject is cut from the University circular for its year. Open a document here if you want the whole thing rather than a single subject.