Artificial Intelligence

Lectures: Tue and Thu 4:00 PM - 5:30 PM. Room: TBD
Instructor: Shashi Prabh
Prerequisites: Introduction to Computation and Programming (CSE 100), Data Structures and Algorithms (CSE 210), Probability and Random Processes (MAT 202), Linear Algebra (MAT 200 or equivalent). Programming ability is a must!

Course description

Artificial Intelligence (AI) is reshaping nearly every domain of human activity, and the rapid rise of foundation models has made first-hand fluency with both classical and modern AI essential. This course builds that foundation. The syllabus covers intelligent agents, problem solving by search (uninformed, informed, local), constraint satisfaction problems, logic and knowledge representation, probabilistic reasoning with Bayesian networks, probabilistic reasoning over time (Markov chains, Hidden Markov Models, forward and Viterbi algorithms), utility theory and decision networks, Markov decision processes, reinforcement learning, machine learning and neural networks, deep learning, transformers and large language models, and AI ethics, safety and alignment.

Course objectives

• Build an in-depth understanding of the fundamentals of Artificial Intelligence and prepare students for more advanced AI courses.
• Develop the ability to represent knowledge and reason under uncertainty.
• Develop problem-solving skills using AI techniques, and the judgement to differentiate between techniques and apply the right one to a given problem.
• Build hands-on competence with widely used AI tools and libraries, and the ability to design and implement (in Python) efficient autonomous agents — including LLM-augmented agents built with modern tooling.
• Develop the ability to evaluate AI applications and their limitations, including the ethical and societal implications of modern AI systems.

Learning outcomes

On successful completion of this course, students will be able to:

• Recognise and characterise artificial intelligence systems in terms of agents, environments and performance measures.
• Formulate problems and apply appropriate search, constraint-satisfaction, logical, probabilistic, and decision-theoretic techniques to solve them.
• Implement core AI algorithms (search, inference, value iteration, basic learning) in Python from first principles.
• Train and evaluate basic neural-network models, and use modern foundation-model tooling (transformers, prompting, retrieval-augmented generation) to build LLM-based agents.
• Critically assess the capabilities, limitations, and ethical implications of AI systems, including bias, hallucination, evaluation pitfalls, and alignment concerns.

Textbook

• AI: A Modern Approach, Russell and Norvig, 4th Edition, Pearson, 2022

Grading

• Assignments: 10%
• Project: 10%
• Quizzes: 10%
• Midterm exam: 35%
• Final exam: 35%

Helpful Advice ( a.k.a. expectation from the students! )
Pay attention and take notes! Get doubts cleared during the lecture itself -- do not hesitate to ask questions in class. Before attending a lecture, review your notes and scan the portion of the textbook that will be covered (see the course calendar page here). Do assignments on your own. If you happen to miss some session(s), do talk to someone else who attended or the TA to find out the topics covered and any announcement made.

Links
Course calendar
Course policies