Course Information

Short Intro

The objective of this course is to provide a complete introduction to Artificial Intelligence.

Description

Technically, there are two general forms of inference to deal with AI problems: logical and probabilistic inference. Agents in the real world need to handle partially known and uncertain worlds, which is best modeled by probabilistic approaches. So modern treatments are primarily based on statistics, and our discussions will also focus on that route.

Teaching Plan

This course is organized into a 12-week session (4 hours per week). The main contents are listed below:

本课程计划安排12周内容，主要讲授内容分别如下：

1. 导言；智能体 Intro to AI, Rational Agents, Ch. 1, 2; State Spaces, Uninformed Search, Ch. 3.1–3.4
2. 搜索；复杂环境 Informed Search: A* and Heuristics, Ch. 3.5–3.6; Local Search Ch. 4.1–4.2
3. 命题逻辑与规划 Propositional Logic and Planning, Ch. 7.1–7.4; Boolean Satisfiability, DPLL, Ch. 8.1–8.2
4. CSP, Ch. 5.1; Ch. 5.2–5.5
5. 博弈论 Games: Trees, Minimax, Pruning, Ch. 6.1–6.3; Games: Expectimax, Monte Carlo Tree Search, Ch. 6.4–6.5
6. Markov Decision Processes: Intro, Ch. 17.1; MDPs: Dynamic Programming, Ch. 17.2
7. Reinforcement Learning, Ch 22.1–22.6; Ch 22.1–22.6
8. Probability Review, Bayesian Networks: Ch. 13.1; Bayes Nets: Syntax and Semantics: Ch. 13.2
9. Bayes Nets: Variable Elimination: Ch. 13.3; Bayes Nets: Sampling: Ch. 13.4
10. Markov Chains, HMMs, filtering, Viterbi: Ch. 14.1–14.5; Dynamic Bayes Nets, Particle Filtering: Ch. 16.1–16.3
11. Utility Theory, Rationality, Decisions: Ch. 16.1–16.3; Decision Networks and VPI: Ch. 16.5–16.7

Schedule

Monday, Wednesday; 9-405.

Evaluation

• Attendance & participation: 20%
• Assignments & exercises: 30%
• Exam: 50%

Textbook

Not mandatory but recommended:

• Russell, Artificial Intelligence: A Modern Approach，即《人工智能: 一种现代方法》

Resources

• Artificial Intelligence: A Modern Approach, 4th US ed¹