Dr. Murat Dundar
Associate Professor
Computer & Information Science Dept.

 

CSCI 590: Machine Learning

Instructor: Murat Dundar
Class time:
1:30 - 2:45pm (TR)
Room:
SL 050
Email: mdundar 'at' iupui.edu

Announcements:

  • Jan 10, 2017: You can download course materials from this site as they become available. Homework solutions will be posted on Canvas.
  • Jan 26, 2017: Assignment 1 posted.
  • February 9, 2017: Assignment 2 posted.
  • February 21, 2017: This year's ML competition has been launched.
  • March 21, 2017: Proposals are due
  • March 27, 2017: Competition is open for test submissions.
  • March 31, 2017: Assignment 3 posted.
  • April 19, 2017: Assignment 4 posted.
  • April 27, 2017: Midterm
  • May 2, 2017: Project presentations
  • May 5, 2017: Competition ends, reports are due.

Syllabus:

You can download the course syllabus here.


Lectures:

  Lecture 1: introduction, fundamental problem in ML slides
 
  Lecture 2: probability theory, distributions, expectations. slides
Video Lecture (recommended): Introduction to Machine Learning by Iain Murray watch
 
  Lecture 3: maximum likelihood estimation, sufficient statistics. slides
Related readings: PRML sections 2.3.4 and 2.4.1
 
  Lecture 4: maximum likelihood estimation, basic concepts in information theory slides
Related readings: PRML sections 2.3.4, 2.4.1, 1.6
 
  Lecture 5: Bayes rule, naive Bayes, Bayesian inference for Multinomial slides
Related readings: PRML sections 2.1, 2.2
 
  Lecture 6: Gaussian partitions, Bayesian inference for Gaussians slides
Related readings: PRML sections 2.3.1, 2.3.2, 2.3.3, 2.3.6, 2.3.7
 
  Lecture 7: Nonparametric methods, K-nearest neighbor classification, linear regression slides
Related readings: PRML sections 2.5, 3.1
 
  Lecture 8: Regularized least squares slides
Related readings: PRML sections 3.1, 3.2
 
  Lecture 9: Bayesian linear regression, Bayesian model comparison, discriminant function slides
Related readings: PRML sections 3.3, 3.4, 4.1
 
  Lecture 10: Least squares for classification, Fisher’s Linear Discriminant slides
Related readings: PRML sections 4.1
 
  Lecture 11: Perceptron algorithm, probabilistic generative models, probabilistic discriminative models slides
Related readings: PRML sections 4.2, 4.3
 
  Lecture 12: Laplace approximation, BIC, Bayesian logistic regression slides
Related readings: PRML sections 4.4, 4.5
 
  Lecture 13: VC Theory and large-margin classifiers slides
Related readings: PRML section 7.1
Recommended Reading: A tutorial on SVMs for pattern recognition pdf
 
  Lecture 14: Support vector machines and KKT theorem slides
 
  Lecture 15: Nonlinear SVM, VC theory revisited, support vector regression slides
Demo materials for SVM svm demo
 
  Lecture 16: Dimensionality Reduction (PCA and SVD) slides
 
  Lecture 17: Latent Semantic Indexing, Latent Dirichlet Allocation, Non-negative Matrix Factorization slides
 
  Lecture 18: Graphical Models (Bayesian networks) slides
Related readings: PRML section 8.1, 8.2, 8.3
 
  Lecture 19: Inference on a chain, factor graphs slides
Related readings: PRML section 8.4.1-3
 
  Lecture 20: Sum-product and Max-sum algorithms for efficient exact inference slides
Related readings: PRML sections 8.4.4-5
 
  Lecture 21: The Expectation-maximization Algorithm slides
Related readings: PRML sections 9.1-9.4
 
  Lecture 22: EM for Gaussian Mixture Models slides
Related readings: PRML sections 9.1-9.4
 
  Lecture 23: Variational Bayes slides
Related readings: PRML sections 10.1
 
  Lecture 24: Expectation Propogation Algorithm slides
Related readings: PRML sections 10.7
 
  Lecture 25: MCMC Tutorial by Iain Murray slides
 
  Lecture 26: MCMC continues slides

Assignments:

Assignment 1: PDF

Assignment 2: PDF  PubMed Dataset

Assignment 3: PDF 

Assignment 4: PDF

 


Project:

Please check the competition website for information about this year's competition.


Grading:

Assignments: 20%
Midterm: 30%
Project Proposal: 5%
Project Presentation: 5%
Project Progress: 5%
Project Report: 15%
Project Results & Relative Rank: 20%


Template for Reports:

  • Download Latex template for project reports. See PDF generated.
  • Download Miktex (the compiler for Latex)
  • Download TexnicCenter (the editor for Latex)

When you configure TexnicCenter you need to choose Miktex as the compiler.