Fall 2016 Artificial Intelligence (CS440/ECE448 Sections Q3, Q4, ONL)

The goal of Artificial Intelligence (AI) is the design of agents that can behave rationally in the real world by sensing their environment, planning their goals, and acting to optimally achieve these goals. This course provides an introductory survey to the techniques and applications of modern AI. The course will cover a broad range of conceptual approaches, from combinatorial search to probabilistic reasoning and machine learning, and a broad range of applications, from natural language understanding to computer vision. Lectures will stress not only the technical concepts themselves, but also the history of ideas behind them. Lectures: Tuesdays and Thursdays, 3:30PM-4:45PM, 1404 Siebel Instructor: Svetlana Lazebnik (slazebni -at- illinois.edu) Office hours (3308 Siebel): TBA Tuesdays and Thursdays 1:30-3PM or by appointment. TAs: Hyo Jin Do (hjdo2), Manav Kedia (mkedia2), Shreya Rajpal (srajpal2), Xuesong Yang (xyang45), Aditi Adhikari (adhikar). TA office hours (207 Siebel): Mondays 5PM-7PM, Wednesdays 5PM-7PM, Fridays 4PM-7PM Contacting the course staff: For emergencies and special circumstances (including extension requests), please email the instructor. For questions about lectures and assignments, use Piazza. For questions about your scores (including regrade requests), email the responsible TAs. Always check announcements and Piazza for short-notice changes to instructor and TA office hours!

Prerequisites: data structures (CS 225 or equivalent), algorithms highly desirable, basic calculus, familiarity with probability concepts a plus but not required.

Textbook: Stuart Russell and Peter Norvig, Artificial Intelligence: A Modern Approach, 3rd edition.

Grading scheme:

• Programming assignments or MPs (mini-projects): 60%.
  • MPs can be done in groups of up to three people. Three-unit students must work with three-unit students and four-unit students must work with four-unit students.
  • Four-unit students will have to solve extra problems on every assignment.
• Midterm and final: 40%.
  • Short answer format, closed book, closed notes, no electronic devices.
• Extra credit:
  • Three-unit students can get extra credit for doing four-unit problems (the raw point value will be discounted by 50%).
  • For four-unit students who drop down to three units, the points earned on four-unit problems will convert to extra credit (with a discount of 50%).
  • A bonus of up to 3% of total course grade will also be given for active participation.

For details, see the grading scheme and statistics from a previous semester.

  Be sure to read the course policies!

Syllabus (tentative)

• AI: Concepts and history
• Solving problems by searching
  • Uninformed search
  • Informed search
  • Constraint satisfaction problems
• Games
  • The minimax principle
  • Modern game-playing systems
  • Game theory
• Planning
• Probabilistic reasoning
  • Basic probability concepts
  • Bayesian inference
  • Naive Bayes models
  • Bayesian networks
• Sequential probabilistic reasoning
  • Markov decision processes
  • Reinforcement learning
• Machine learning
  • Supervised vs. unsupervised learning
  • Decision trees
  • Nearest neighbor classifiers
  • Neural networks
  • Deep learning
• Applications of AI
  • Robotics
  • Computer vision
  • Natural language processing

Announcements

• November 30: Lana's office hours next Tuesday will be moved to 1-2:30PM.
• November 29: Conflict exam requests are due by the end of Thursday, December 1st. please email Lana Lazebnik and Aditi Adhikari (adhikar -at- illinois.edu) with the details of your conflict (including the dates you are unavailable).
• November 29: Assignment 4 Part 1 instructions have been updated with additional details and clarifications.
• November 29: Lecture is cancelled today due to an investiture ceremony being held in the classroom by the CS department.
• November 15: Assignment 4 is out, due Monday, December 5, 11:59:59PM.
• October 27: Assignment 3 is out, due Monday, November 14, 11:59:59PM.
• October 25: Wednesday TA office hours will now be 5-7PM.
• October 25: The final exam will be on Monday, December 12th, 9-10:15AM, in 1404 Siebel. If you have a conflict for this time, please email Lana Lazebnik and Aditi Adhikari (adhikar -at- illinois.edu) with the details of the conflict (including the dates you are unavailable).
• October 11: Solutions to practice midterm problems have been posted on Compass.
• October 11: Lana is traveling this week and her office hours are cancelled. The review session today will be led by Prof. Mark Hasegawa-Johnson.
• October 7: Assignment 2 instructions have been updated to provide clarifications about offensive vs. defensive evaluation functions and extended rules in Part 2.
• October 4: Midterms will take place in class next Thursday, October 13th. If you need to request a conflict midterm, please email Lana Lazebnik and Aditi Adhikari (adhikar -at- illinois.edu) by the end of tomorrow (Wednesday, October 5th). In your email, please give the details of your conflict, including the dates on which you will be unavailable to take the midterm.
• September 29: Assignment 2 is out, due Monday, October 24, 11:59:59PM.
• September 22: For Assignment 1 Part 1.2, please download new smallSearch and mediumSearch files, which have been made easier (there are fewer dots).
• September 21: If you are not yet registered for the class and are not working in a group with somebody who is registered, please email the instructor ASAP to be added to Compass so you can submit Assignment 1.
• September 21: The original cube3_1 input in Part 2 of Assignment 1 was not solvable, so we have replaced it by a different input.
• September 8: Assignment 1 is out, due Monday, September 26, 11:59:59PM.
• September 6: Lana's office hours will be cancelled on Thursday, September 8th, and Thursday, September 15.
• September 1: We are creating a wait list for both sections of this class. If you wish to get on the wait list, please email slazebni -at- illinois.edu with the subject "CS440/ECE448 wait list" and one line in the body of the email formatted as follows: Lastname, Firstname (netid), section preference. For example: Smith, John (jsmith), Q3, R3. You will receive an acknowledgment with your number on the wait list, and will be registered for the first available section as spots open up. Wait list requests will only be accepted through noon on Friday, September 2nd.
• August 30: TA and instructor office hours have been posted above. They will start on Wednesday, August 31.
• August 16: Welcome! Please enroll yourself in the Piazza discussion board. Participation on the board is voluntary, but it can help you to earn participation bonus points.

Schedule (tentative)

Date	Topic	Readings and assignments
August 23	Intro to AI: PPT, PDF	Reading: Ch. 1
August 25	History and themes: PPT, PDF	Reading: Ch. 1
August 30	Agents: PPT, PDF	Reading: Ch. 2
September 1	Search intro: PPT, PDF	Reading: Ch. 3
September 6	Uninformed search: PPT, PDF	Reading: Sec. 3.1-3.4
September 8	Informed search: PPT, PDF	Reading: Sec. 3.5-3.6 Homework: Assignment 1 is out
September 13	Constraint satisfaction problems: PPT, PDF	Reading: Ch. 6
September 15	CSPs cont. (slides above)
September 20	Minimax search: PPT, PDF	Reading: Ch. 5
September 22	Stochastic tree search and stochastic games: PPT, PDF	Assignment 1 due September 26 11:59:59PM
September 27	Game theory: PPT, PDF	Reading: Sec. 17.5-17.6
September 29	Game theory cont.	Homework: Assignment 2 is out
October 4	Planning: PPT, PDF	Reading: Ch. 10
October 6	Probability: PPT, PDF	Reading: Ch. 13
October 11	Midterm review: PDF
October 13	Midterm (in class)
October 18	Probability cont.	Reading: Ch. 13
October 20	Bayesian inference: PPT, PDF	Assignment 2 due October 24 11:59:59PM
October 25	Bayesian networks: PPT, PDF	Reading: Ch. 14
October 27	Bayesian networks cont.	Homework: Assignment 3 is out
November 1	Bayesian network inference: PPT, PDF	Reading: Ch. 20
November 3	Hidden Markov models: PPT, PDF	Reading: Ch. 15, sec. 23.5
November 8	Markov decision processes: PPT, PDF	Reading: Ch. 17
November 10	Reinforcement learning: PPT, PDF	Reading: Ch. 21 Assignment 3 due November 14
November 15	Machine learning: PPT, PDF	Homework: Assignment 4 is out
November 17	Perceptrons, neural networks: PPT, PDF
November 29	Class cancelled
December 1	Deep learning: PPT, PDF	Assignment 4 due December 5
December 6	Final review: PDF
December 12	Final exam: 9-10:15AM, 1404 Siebel