Computer Science 246/446

Mathematical Foundations of Artificial Intelligence

Spring 2013

Instructor: Dan Gildea office hours Tu/Th 2-3pm
TA: Iftekhar Naim office hours M/W 2:30-3:25pm
Location: M/W 3:25-4:40pm, Dewey 2110E

Homeworks

Lecture Notes

  • Vector calculus review: 3:25pm Fri 1/18, CSB 632
  • Python help session: 5pm Tue 1/22, CSB 601

Required text: Kevin P. Murphy, Machine Learning: A Probabilistic Perspective

The following are useful references in addition to the reading material assigned for each class:

  • Christopher M. Bishop, Pattern Recognition and Machine Learning.
  • Stuart Russell and Peter Norvig, Artificial Intelligence, A Modern Approach.
  • Larry Wasserman, All of Statistics, 2004.
Readings not taken from Murphy are available in the box in the CS mailroom.

Syllabus

Onwe will coverwhich means that after class you will understandif before class you have read
1/16 Probability Theory independence, bayes rule murphy ch 1
1/23 Information Theory entropy, kl-distance, coding murphy ch 2
1/28 Probabilistic Inference priors on continuous variables murphy 3.2
1/30 Probabilistic Inference priors: bayesian reasoning, MAP murphy 3.4
2/4 Perceptron stochastic gradient descent murphy 8.5
2/6 Backpropagation DP for gradient descent murphy 16.5
2/11 Support Vectors strong duality murphy 14.5
2/13 Support Vectors the kernel trick murphy 14.4
2/18 Probabilistic Inference message passing murphy 20.2
2/20 Tree decomposition cyclic graphs murphy 20.4
2/25 Expectation Maximization L = Q + H + D murphy 11.4
2/27 Expectation Maximization mixture of gaussians murphy 11.4
3/4 Sampling Markov Chain Monte Carlo murphy 24.3
3/6 Sampling Gibbs Sampling murphy 24.2
3/18 Learning Theory Error bounds kearns and vazirani ch 1, 3
3/20 Review
3/25 Midterm
3/27 Midterm Solutions
4/1 Eigenvectors least squares, PCA murphy 12.2
4/3 Logistic Regression maximum entropy murphy 8.3
4/8 Optimization LBFGS nocedal 8.1, 9.1
4/10 Hidden Markov Models forward-backward murphy 17.3, 17.4
4/15 Particle Filters slam murphy 23.5
4/17 Reinforcement Learning q-learning sutton ch 3, 4.3, 4.4, 6.1, 6.5, 7.2, 11.1
4/22 Sparse Linear Models lasso murphy 13.3, 13.4
4/24 Sparse Linear Models sparse coding murphy 13.8, lee
4/29 Uniform Deviation Inequalities Daniel Stefankovic devroye ch3
5/1 Review come to class with questions!
Final Exam: Monday May 6, 8:30-11:30am, Dewey 2110E.

Grading

  • Final exam: 35%
  • Homeworks: 30%
  • Midterm: 25%
  • NB: 5%
  • Class participation: 5%

gildea @ cs rochester edu
May 3, 2013