Probabilistic Foundations of Artificial Intelligence

Overview

How can we build systems that perform well in uncertain environments and unforeseen situations? How can we develop systems that exhibit ''intelligent'' behavior, without prescribing explicit rules? How can we build systems that learn from experience in order to improve their performance? We will study core modeling techniques and algorithms from statistics, optimization, planning, and control and study applications in areas such as sensor networks, robotics, and the Internet. The course is designed for upper-level undergraduate and graduate students.

Topics covered

• Tutorial in search (BFS, DFS, A*), constraint satisfaction and optimization
• Tutorial in logic (propositional, first-order)
• Probability
• Bayesian Networks (models, exact and approximate inference, learning)
• Temporal models (Hidden Markov Models, Dynamic Bayesian Networks)
• Probabilistic planning (MDPs, POMDPs)
• Reinforcement learning
• Combining logic and probability

News

• PAI final exam will be on Tuesday (Jan 28), from 09:00-11:00, in HIL G 15.
• The review (Q&A) session will be on Tuesday, January 21th, from 3-4pm in ML F 34. Please send your questions to (all the) TAs until Jan. 19th.
• Past years exams (accessible only from ETH domain) have been posted.
• As noted on the VVZ, the lecture room has been changed to CHN C 14. The exercises will remain in HG E 41.
• The lecture room will likely change due to the class size. Please stay tuned.
• There will be a second recitation session Friday 14-15. Please attend according to the initial of your last name.

Details

• VVZ Information: See here.
• Lecture: Friday 10-12 in CHN C 14
• Recitations: Friday 13-14 (Last names A-L) and 14-15 (Last names M-Z) in HG E 41
• Teaching assistants:
  • Hastagiri Vanchinathan [hastagiri (at) inf (dot) ethz (dot) ch]
  • Yuxin Chen [yuxin (dot) chen (at) inf (dot) ethz (dot) ch]
  • Baharan Mirzasoleiman [baharanm (at) inf (dot) ethz (dot) ch]
• Textbook: S. Russell, P. Norvig. Artificial Intelligence: A Modern Approach (3rd Edition).

Homeworks

• Homework 1 [pdf]. Solutions 1 [pdf]
• Homework 2 [pdf]. Solutions 2 [pdf]
• Homework 3 [pdf]. Solutions 3 [zip]
• Homework 4 [pdf]. Supplementary files [zip]. Solutions 4 [pdf]
• Homework 5 [pdf]. Solutions 5 [pdf], Supplementary [zip]
• Homework 6 [pdf]. Data files: [trainingData.txt] and [testingData.txt]. Solutions 6 [pdf]

Lecture Notes

• Sept 20: Introduction, Uninformed search (Chapters 1 and 2)
• Sept 27: Informed search (Chapters 3.1-3.6)
• Oct 4: A primer in Logic (Chapters 7.1-8.3)
• Oct 11: First order logic; (Chapters 9.1-9.2, 9.5)
• Oct 18: Intro to probability (Chapters 9.1-9.2, 9.5, 13.1-5)
• Oct 25: Bayesian Networks (Chapters 14.1-14.2); JavaBayes applet
• Nov 1: Inference in Bayesian Networks (Chapters 14.4)
• Nov 8: Approximate Inference; Information Gathering (Chapter 14.5; Bishop Chapter 8.4)
• Nov 15: Temporal models: HMMs, Kalman Filters, DBNs (Chapter 15.1-15.5)
• Nov 22: Probabilistic Planning: MDPs (Chapter 17.1-17.3). Value iteration demo (at UBC)
• Nov 29: Learning Bayesian Networks (Chapter 20.1-2).
• Dec 13: Reinforcement learning (Chapter 21.1-3). Q-learning demo (at Northwestern).
• Dec 20: Logic and probability (Chapter 14.6). Alchemy (University of Washington).

Recitations

• Sept 27: Search [pdf]
• Oct 4: Propositional Logic [pdf], HW1 Solution
• Oct 11: First-order Logic [pdf]
• Oct 18: Probability [pdf]
• Oct 25: HW2 Solution
• Nov 1: Bayesian Networks [pdf]
• Nov 8: HW3 Solution [pdf]
• Nov 15: Inference [pdf]
• Nov 22: HW4 Solution, Plobablistic Planning [pdf]
• Nov 29: Learning Bayes Nets [pdf]
• Dec 13: HW5 Solution
• Dec 20: Q-learning, Bandits, HW6 Solution [pdf]

Old Exams

• PAI exam, 2011 fall [pdf]
• PAI exam, 2012 fall [pdf]

Relevant Readings

• Christopher M. Bishop. Pattern Recognition and Machine Learning. Springer, 2007 (optional)