Deep and Reinforcement Learning 2026

MSc Data Science and MBA Joint Course at MSE

Logistics

Timings: Monday–Thursday, 11:15 AM–12:45 PM
Venue: Cognizant Lab, COE Building, second floor
Teaching Assistant: Rajat Kumar Mishra
Students: MSc Data Science (Semester 3, 2026) and MBA (joint)

Syllabus

1. Foundations of Neural Networks
Multilayer perceptrons, feedforward networks, activation functions, loss functions, empirical risk minimization, gradient descent, backpropagation, regularization.

2. Architectures for Representation Learning
Convolutional neural networks, filters, pooling, transfer learning, recurrent neural networks, vanishing and exploding gradients, LSTM networks, autoencoders.

3. Generative Models and Introduction to Reinforcement Learning
Probabilistic models, latent representations, Boltzmann machines, restricted Boltzmann machines, deep belief networks, sequential decision problems, states and actions, rewards and returns, policies and value functions.

4. Markov Decision Processes and Dynamic Programming
Markov chains, MDPs, transition probabilities, reward functions, discounting, Bellman principle of optimality, policy evaluation, policy improvement, value iteration.

5. Approximate Methods in Reinforcement Learning
Monte Carlo prediction, action-value estimation, Monte Carlo control, exploration and exploitation, importance sampling, value-function approximation, stochastic gradient methods, feature construction, least-squares methods.

References

  1. Ian Goodfellow, Yoshua Bengio, Aaron Courville, Deep Learning, MIT Press, 2016.
  2. Richard S. Sutton, Andrew G. Barto, Reinforcement Learning: An Introduction, MIT Press, 2nd edition, 2018.

Assignments

Tests