Course overview

This course introduces students to artificial intelligence, providing the fundamental principles and essential techniques for analyzing and solving complex real-world problems. The material spans machine learning, neural networks, language modeling, intelligent search, reinforcement learning, and reasoning. By the end of the semester, students will have gained a broad set of tools and models that form the basis for AI applications and practices.

The course emphasizes both basic concepts relevant to Artificial Intelligence and also the practical know-hows. Students will practice through mathematical and programming assignments to study Artificial Intelligence, project proposal and final project presentations. The tentative list of topics includes:

• Learning: Basics of supervised machine learning, neural networks, as well as their applications in AI such as large language models.
• Intelligent search and games: search algorithms such as A-star search, games such as zero-sum games, Nash equilibrium.
• Sequential decision-making and reasoning: Markov decision processes, Bellman equation, and reinforcement learning automated reasoning, uncertainty, and Bayesian networks

Prerequisites: We will use various aspects of linear algebra and probability when needed during the course. When we talk about a specific concept, we will cover the ground up. We will use python as the main programming language when it comes to coding exercises. We will provide detailed instructions when a relevant tool is required as part of a homework problem.

There will be 6 homeworks, plus a take-home exam as part of the course work. In addition, we will do a course project that can be done in teams of students. We will provide you with detailed materials and case studies as part of the course project handouts.

Course syllabus

Week 1, Sep 4: Course introduction

• Course plan, foundation and history of Artificial Intelligence

Week 2, Sep 8: Supervised learning I, Sep 11: Supervised learning II

• Regression and classification, including examples of supervised learning tasks

• Linear classifiers

• Loss function, optimization, and robustness

Week 3, Sep 15: Neural networks I, Sep 18: Neural networks II

• Neural network architecture and learning algorithms

• Convolutional neural network

Week 4, Sep 22: Neural networks III, Sep 25: Backpropagation and regularization

• PyTorch implementations.

• Backpropagation and regularization

Week 5, Sep 29: Language models I, Oct 2: Language models II

• Basic concepts of language modeling, including adaptation and prompting

• Next-word prediction

Week 6, Oct 6: Implementation of language models in PyTorch, Oct 9: An example with sentiment analysis

• Training and adaptation with PyTorch implementations.

• The naive Bayes classifier.

• Basics of logistic regression.

Week 7, Oct 16: Intelligent search I

• Breadth-first search

• Depth-first search

• Uniform cost search

Week 8, Oct 20: Intelligent search II

• Informed search algorithms: greedy best-first search

• A-star search

• Beam search

Week 8, Nov 3: Search and games I, Nov 6: Games II

• Minimax game and speed up search

• Zero-sum game and None-zero-sum game

Week 9, Nov 10: Sequential decision-making and Markov decision processes I, Nov 13: Sequential decision-making and Markov decision processes II

• Stocastic search problems and Markov decision processes

• Reinforcement learning

Week 9, Nov 17: Bayesnets---Coping with uncertainty

• Probabilistic inference and Bayesian networks

Course information

You are responsible for keeping up with all announcements made in class and for all changes in the schedule that are posted on the Canvas website.

The grade will be based on the following:

• Homeworks: 40%

• Exam (take-home): 40%

• In-class project presentations and participation: 20%

Textbooks for reference:

• Artificial Intelligence: A Modern Approach, 4th US ed. Stuart Russell and Peter Norvig.