AIML Summer Course — IIT Bombay CSE

Welcome. The first lecture slide is released on Monday, June 1, at 00:00 hours. All enrolled students must register on ASC and read the slides shared every week to stay in sync with the course. Weekly slides will be shared in the beginning of each week. Please go over them carefully and read the backup material from the book or the content of the original course (see below).

Programming Assignment will be administered in-person at the end of Week 4. Venue and timing will be announced. Bring your laptop.

End Semester Exam will be held in-person (closed book, with one formula sheet) at the end of Week 6. Date and venue TBD.

Course Overview

This is a condensed 6-week summer course on Artificial Intelligence and Machine Learning, designed to give students a solid foundation across the core topics: supervised learning, deep neural networks (including Transformers), classical AI search, and multi-agent AI (game theory). The course is self-contained and targets students who want a focused, accessible introduction to modern AI/ML.

The course draws from the material of CS 217+240 (2024). Questions are designed at easy-to-medium difficulty, suitable for students building foundational understanding.

Piazza

We will have the course-related discussions on Piazza. Here is the course link.

Prerequisites

Linear algebra: vectors, matrices, eigenvalues
Calculus: partial derivatives, chain rule
Probability and statistics: distributions, expectations, MLE
Basic Python programming (numpy, matplotlib)

Lectures

Week	Title	Topics Covered	Materials
Wk 1	Optimization & Linear Regression	What is Machine Learning? Supervised learning Convexity, gradient descent (batch, SGD, mini-batch) Linear regression, normal equation, MLE Basis functions, overfitting, ridge regularization	Slides
Wk 2	Classification & SVM	Binary classification, logistic regression, sigmoid Cross-entropy loss, multiclass (softmax) Evaluation: precision, recall, F1 Hard/soft-margin SVM, kernel trick	Slides
Wk 3	Neural Networks I	Perceptron, MLP, activation functions (ReLU, GELU) Universal approximation theorem Backpropagation, chain rule, vanishing gradients He/Xavier initialization, Adam optimizer, dropout, batch norm	Slides
Wk 4	Neural Networks II	CNN: convolution, pooling, skip connections (ResNet) RNN, LSTM, GRU Attention mechanism (scaled dot-product, multi-head) Transformer architecture, BERT vs GPT, LLMs	Slides
Wk 5	Classical AI Search	Search problem formulation (state, actions, cost) BFS, DFS, IDDFS, Uniform Cost Search Heuristics, admissibility, A* algorithm Optimality proof for A*, local search	Slides
Wk 6	Multi-Agent AI	Normal form games, dominant strategies, Nash equilibrium Mixed strategies, Nash's theorem, zero-sum games Minimax theorem, game trees, minimax algorithm Alpha-beta pruning, MCTS, AlphaGo	Slides

Evaluation

Component	Marks	When	Mode
Programming Test Regression, Classification, MLP — covers Weeks 1–4	50	End of Week 4: July 1, 2-4 PM	In-person, proctored, 2 hrs
Written End-Semester Exam Covers all 6 weeks; closed book + 1 formula sheet	50	End of Week 6: July 15, 2-4 PM	In-person, proctored, 2 hrs
Total	100	—	—

Academic Integrity. Students are expected to uphold the highest standards of academic honesty. Copying in examinations or unauthorized collaboration will be dealt with strictly in accordance with IIT Bombay's procedures for academic malpractice.