CSE 5522 Artificial Intelligence II: Advanced Techniques

Advanced concepts, techniques, and applications of artificial intelligence, including knowledge representation, learning, natural language understanding, and vision.

Details
Textbook:
Grading

Participation (5%)

You will receive credit for engaging in class discussion, asking and answering questions related to the homework on Piazza online discussion board.

Homeworks (15%)

Written homeworks will be very short (one or two exam-style questions) and will be graded in a good/mediocre/incomplete basis. You should be prepared to do regular work each week to keep up with the material and the assignments. Homeworks due before class on day X will include topics we will discuss in class on day X. We will talk about solution in class if people have questions. Homework assignments may NOT be turned in late. Homeworks are NOT accepted by email. There will be 1 grace homework grade per semester, that is, each student receiving full credit for the lowest or a missing homework grade.

Projects (30%)

Programming projects will be in Python, and should be submitted to Carmen by 11:59pm on the day it is due (unless otherwise instructed). Each student will have 3 flexible days to turn in late programming projects throughout the semester. As an example, you could turn in the first project 2 days late and the second project 1 day late without any penalty. After that you will loose 20% for each day the project submission is late. Please email your code to the instructor in case there are any technical issues with submission.

Midterm (20%)

Midterm exam will be close book and notes.

Final Exam (30%)

Final exam will be close book and notes.

Grading Scale: Numerical grades will be mapped to letter grades using the standard OSU policy: 93-100 (A), 90-92.9 (A-), 87-89.9 (B+), 83-86.9 (B), 80-82.9 (B-), 77-79.9 (C+), 73-76.9 (C), 70-72.9 (C-), 67-69.9 (D+), 60-66.9 (D), below 60 (E). These cutoffs represent grade minimums. We may adjust grades upward based on class grade distribution curve.

Regrade Policy: If you believe an error has been made in the grading of your exam, you may resubmit it for a regrade - submit a detailed explanation of which problems you think we marked incorrectly and why. Because we will examine your entire submission in detail, your grade can go up or down as a result of a regrade request.

Drop or Withdraw: detailed OSU policy and instructions here

Resources
  • Piazza (QA, discussion, and announcements)
  • Carmen (project submission and restricted resources)
    •
    Academic Integrity
    Any assignment or exam that you hand in must be your own work (with the exception of group projects). However, talking with others to better understand the material is strongly encouraged. Copying a solution or letting someone copy your solution is cheating. Everything you hand in must be your own words. Code you hand in must be written by you, with the exception of any code provided as part of the assignment. MOSS (Measure of Software Similarity) will be used routinely to detect plagiarism on programming assignments. Any collaboration during an exam is considered cheating. Any student who is caught cheating will be reported to the Committee on Academic Misconduct. Please don't take a chance - if you are having trouble understanding the material, let us know (asking on Piazza, in class or during office hours), and we will be happy to help.
    Programming Projects
  • Project 0 - Unix/Python/Autograder Tutorial (optional)
  • Project 1 - Search (optional)
  • Project 2 - Multi-Agent Search (optional)
  • Project 3 - Reinforcement Learning (due 10/9)
  • Project 4 - Ghostbusters (due 11/9)
    • Homework Assignments (subject to change as the autumn 2018 term progresses.)
  • Homework 0 - Math and Python Review (due 8/29)
  • Homework 1a - A* Graph vs. Tree Search (optional)
  • Homework 1 - A* Search and Minimax (due 9/7)
  • Homework 2 - Markov Decision Processes (due 9/21)
  • Homework 3 - Reinforcement Learning: Q-learning (due 9/28)
  • Homework 4 - Reinforcement Learning: Feature-based Representations (optional; up to 2% bonus, if submit in-class 10/5 or 10/10)
  • Homework 5 - Probability Review (optional)
  • Homework 6 - Hidden Markov Models (due 10/31)
  • Homework 7 - Bayes Nets (due 11/16)
  • new Homework 8 - Naive Bayes (optional)

  • Interactive Practice - Alpha-Beta Pruning
  • Example Exam - MDPs, Reinforcement Learning, Markov Model, HMM, Bayes Nets
    •
    Anonymous Feedback
    https://goo.gl/forms/iV3rOU6MzarlxXEr1
    Schedule (subject to change as the autumn 2018 term progresses.)
    Date Topic Required Reading Suggested Reading
    8/22 Course Overview, Reinforcement Learning in NLP Research - Guest Lecture by Alan Ritter 3.1-3.4 Microsoft Ms. PacMan
    8/24 Search Review 3.5, 3.6 Pancake Sorting
    8/29 Game Playing 1 - Minimax (HW0 due) 5.1, 5.2, 5.3, 5.4 How Checkers was Solved
    8/31 Game Playing 2 - Expectimax and Utilities 5.5, 13.1, 13.2, 16.1, 16.2, 16.3 Rosen's note on Alpha Beta Pruning
    9/5 Utilities (cont') and homework recitation
    Reinforcement Learning
    9/7 Reinforcement Learning 1 - Markov Decision Processes (HW1 due) 17.1, 17.2
    9/12 Reinforcement Learning 2 - Value Iteration and Policy Iteration 17.3 Real-life examples of Markov Decision Processes
    9/14 drop deadline
    9/14 Policy Iteration (cont') system of linear equations, dynamic programming
    9/19 Reinforcement Learning 3 - Temporal Difference Learning 21.1, 21.2, 21.3 Google AlphaGo
    9/21 Reinforcement Learning 4 - Q-Learning (HW2 due) 21.1, 21.2, 21.3 Reinforcement learning for autonomous helicopter flying
    9/26 Probability Review (Engineering Expo) 13.3, 13.4, 13.5, 13.6 Andrew Moore's tutorial and Jean Walrand's note on Probability
    9/28 Reinforcement Learning 5 - Function Approximation (HW3 due) 21.4 Sutten & Barto's reinforcement learning book (Ch. 5, 6, 13)
    10/3 Reinforcement Learning 6 - Deep Q-Learning and Policy Gradient 21.5, 21.6 Deep Q-Learning
    Reasoning Under Uncertainty
    10/5 Midterm Review, Markov Models 15.1, 15.2
    10/9 P3 due
    10/10 Hidden Markov Models 1 - Forward algorithm and robot localization 15.1, 15.2
    10/12 no class (fall break)
    10/17 Midterm (in class - Search, Game, Utilities, MDPs, RL)
    10/19 Hidden Markov Models 2 - Particle filtering [lecture notes] 15.3, 15.5
    10/24 Hidden Markov Models (cont') and Speech Recognition 15.3, 15.5 Haytham Fayek's tutorial on Speech Processing
    10/26 Bayes Nets 1 - Probabilistic Representations 14.1, 14.2
    10/26 withdraw deadline
    10/31 NLP Research - Guest Lecture by Micha Elsner (HW6 due)
    11/2 Speech Research - Guest Lecture by Peter Plantinga
    11/7 Bayes Nets 2 - D-Seperation [lecture notes] 14.1, 14.2 Olivier Chapelle's talk on Bayesian Network Click Model for Web Search
    11/9 Bayes Nets 3 - Inference [lecture notes] (P4 due) 14.4 Stanford CS228: Probabilistic Graphical Models
    11/14 Bayes Nets 4 - Sampling 14.5 David Blei's talk on Topic Models and User Behavior
    Machine Learning and Special Topics
    11/16 Naive Bayes (HW7 due) 20.1, 20.2
    11/21 no class (thanksgiving)
    11/23 no class (Columbus day)
    11/28 Naive Bayes (cont') [lecture notes] Text Classification using Naive Bayes by Hiroshi Shimodaira
    11/30 Computer Vision 1 - Applications 18.6 Facebook Accessibility
    12/5 Final Review, Computer Vision 2 - Deep Neural Networks and Visualization 20.1, 20.2.1, 20.2.2 Google DeepDream
    12/12 12:00-1:45pm Final Exam (close book and notes)
    9/5 AI Seminar by Wuwei Lan, 4pm in Dreese 263 (optional) Neural Networks for Sentence Pair Modeling
    9/18 Guest Speaker - Eunsol Choi, 4pm in Dreese 480 (optional) Ultra-fine Entity Typing
    9/19 Guest Speaker - Mark Yatskar, 4:30pm in Dreese 480 (optional)
    11/7 AI Seminar by Mounica Maddela, 4pm in Dreese 263 (optional) A Word-Complexity Lexicon and A Neural Readability Ranking Model for Lexical Simplification