Intro to AI

Description

Schedule and Readings

Module 1. AI and agents.

Lectures:

• Sep 4. Introduction. Lecture Notes.
• Sep 6. Agents. Lecture Notes.
• Sep 11. Symbols, knowledge and reactive agents. Lecture Notes. Recitation.
• Sep 13. State, control and behavior-based agents. Lecture Notes.
• Sep 18. Combining behaviors: subsumption and flocking. Lecture Notes Recitation.

Readings:

• Russell and Norvig. Artificial Intelligence: A Modern Approach. 3rd Edition. Chapter 2, pages 34-61. Scanned copy
• Bourg and Seemann. AI for Game Developers. Chapter 2, Chasing and Evading, pages 6-26. Scanned copy
• Bourg and Seemann. AI for Game Developers. Chapter 4, Flocking, pages 52-79. Scanned copy
• Bourg and Seemann. AI for Game Developers. Chapter 9, Finite State Machines, pages 165-187. Scanned copy
• Braitenberg. Vehicles: Experiments in Synthetic Psychology. MIT 1984. Chapters 1-5, pages 1-25. Scanned copy
• Reynolds. Flocks, Herds, and Schools: A Distributed Behavioral Model. Proceedings of SIGGRAPH 1987. Electronic copy

Projects:

• Mazes Maze Solving. Released: Tue Sep 11. Due date: Tue Sep 25, 3pm. Skeleton Code, Sample Inputs, and Target Behavior. slight correction to trial and error spec 9/14.
• Flockers Exploring behaviors. Released: Tue Sep 25. Due date: Tue Oct 9, 3pm. Skeleton Code, Sample Inputs, and Target Behavior.

Module 2. Problem solving and search.

Lectures:

• Sep 20. More on flocking. Search: World models and problem solving. Lecture Notes
• Sep 25. Search: Combinatorics and methodology. Lecture Notes Recitation.
• Sep 27. Search: Simple strategies and representations. Lecture Notes
• Oct 2. Search, memory and analysis. Lecture Notes
• Oct 4. More examples. Lecture Notes

Readings:

• Russell and Norvig. Artificial Intelligence: A Modern Approach. 3rd Edition. Chapter 3, pages 64-112. Scanned copy
• Russell and Norvig. Artificial Intelligence: A Modern Approach. 3rd Edition. From Chapter 4, pages 120-129. Scanned copy
• Bourg and Seeman. AI for Game Developers. Chapter 6, A* Pathfinding, pages 126-148. Scanned copy

Demos:

• A* pathfinding search. Interactive demo

Projects:

• Search Out: Mon Oct 15. Due: Th Nov 1, 3pm. Th Nov 8, 3pm. Skeleton Code, Sample Inputs. (See description for examples of output.)

Module 3. Perception and learning.

Lectures:

• Oct 9. A* and Best-first search. Lecture Notes
• Oct 11. Introduction to learning. Lecture Notes
• Oct 16. Nearest neighbor. Lecture Notes Midterm Review.
• Oct 18. Midterm Practice Problems Practice Problems with Answers
• Oct 23. Probabilistic classification. Lecture Notes
• Oct 25. Naive Bayes. Lecture Notes
• Oct 30. (Sandy)
• Nov 1. (Sandy)
• Nov 6. Linear classifiers. Lecture Notes
• Nov 8. Similarity, recommendation. Lecture Notes
• Nov 13. Using learning in practice. Lecture Notes

Readings:

• Segaran. Programming Collective Intelligence. Chapter 6: Document Filtering, pages 117-141. Scanned Copy
• Russell and Norvig. Artificial Intelligence: A Modern Approach, 3rd Edition. From Chapter 13: Quantifying Uncertainty, pages 480-499. Scanned Copy
• Segaran. Programming Collective Intelligence. Chapter 2: Making Recommendations, pages 7-28. Scanned Copy

Supplementary material:

• Russell and Norvig. Artificial Intelligence: A Modern Approach, 3rd Edition. Chapter 18: Learning from Examples, pages 683-758. Scanned Copy

Demos:

• K-means clustering
• Probabilistic clustering

Projects:

• Reco. Project 4. Released Th Nov 8. Due date: Tue Nov 20, 3pm.

Module 4. Intelligent action.

Lectures:

• Nov 15. Utility and uncertain decisions. Lecture Notes
• Nov 20. Decisions over time. Lecture Notes
• Nov 27. Dynamic programming. Lecture Notes
• Nov 29. Reinforcement learning. Lecture Notes
• Dec 4. Game-playing and strategy. Lecture Notes

Readings:

• Russell and Norvig. Artificial Intelligence: A Modern Approach, 3rd Edition. Chapter 16: Making Simple Decisions, pages 610-626. Scanned Copy
• Stone. Agents in the Real World. Chapter written for Rutgers Collection, "What is Cognitive Science", 2001. PDF

Demos:

• Blackjack player with reinforcement learning. Applet
• Demo of dynamic programming for uncertain path planning. Applet

Projects:

• RL Reinforcement learning. Released: Tue Nov 27. Due date: Tu Dec 11.

Module 5. AI in context.

Lectures:

• Dec 6. The limits of current AI. Lecture Notes
• Dec 11. AI, science and culture. Lecture Notes

Readings:

• Russell and Norvig. Artificial Intelligence: A Modern Approach, 3rd Edition. Chapter 26: Philosophical Foundations, pages 1020-1040. Scanned Copy

Supplementary material:

• William Gibson, Neuromancer, Ace Books. 1984. Text

Further information:

• Daniela Rus's robot pebbles Project page.
• IBM's Watson Jeopardy computer Project page and New York Times applet. Technical Paper
• Stanley robot winner of the DARPA grand challenge. Project page and Good interview with project member
• ICT's virtual humans. Project page and Technical Paper

• Dec 14, 8am-11am. Final exam. Practice Problems and Solutions from last year. For this year, link is Practice Problems and Solutions.