CSC2535 Spring 2011 - Lectures

DRAFT LECTURE SCHEDULE: SUBJECT TO CHANGE

• January 12
  Lecture 1: Overview of Machine Learning and Graphical Models
  notes as ppt , notes as .pdf
  Reading: Bishop, Chapter 8: pages 359-418
  

• January 19
  Lecture 2: Approximate Inference and Learning in Directed Graphical Models
  notes as ppt , notes as .pdf
  Reading: Bishop, Chapter 10: pp 461-505
  

• January 26 (first assignment posted on web)
  Lecture 3a: Approximate Learning Methods for Energy-Based Models
  notes as ppt , notes as .pdf
  Lecture 3b: Learning Deep Boltzmann Machines
  notes as .odp , notes as .pdf
  Required reading: Training Products of Experts by Minimizing Contrastive Divergence. [pdf]
  Required reading: Training Restricted Boltzmann Machines using Approximations to the Likelihood Gradient. [pdf]
  Required reading: Estimation of non-normalized statistical models using score matching. [pdf]
  

• February 2
  Lecture 4: Methods for Sampling from Distributions
  notes as .ppt , notes as .pdf
  Reading: Bishop, Chapter 11: pages 523-558
  

• February 9: First assignment due (at start of class)
  Lecture 5a: More ways to fit energy-based models
  notes as ppt , notes as .pdf
  Lecture 5b: Object Recognition and Information Retrieval in Deep Belief Nets
  notes as ppt , notes as .pdf
  Required reading: Reducing the dimensionality of data with neural networks. [pdf]
  Required reading: Small codes and large image databases for recognition.
  [pdf]
  Required reading: Using Very Deep Autoencoders for Content-Based Image Retrieval
  [pdf]
  Optional reading: A Fast Learning Algorithm for Deep Belief Nets. [pdf]
  

• February 16 (second assignment posted on web)
  Lecture 6a: Learning Three-Way Interactions for Image Transformations and Animate Motion.
  notes as .ppt , notes as .pdf
  Required reading: Learning to represent spatial transformations with factored higher-order Boltzmann machines. [pdf]
  Required reading: Factored Conditional Restricted Boltzmann Machines for Modeling Motion Style. [pdf]
  Lecture 6b: Modeling character strings with recurrent neural networks
  notes as .pdf
  Required reading: Hardcopy to be handed out in class.

• February 23 No Lecture (reading week)
  

• March 2 Second assignment due (at start of class)
  Lecture 7: Models of words and documents
  Part 1 as .pdf Part 2 as .pdf
  Required reading: Latent Dirichlet allocation. [html]
  

• March 9
  Lecture 8: Generative models for language, speech, and static images.
  Language models: notes as .pdf
  .....Required reading: [.pdf]
  Speech recognition and image patch modeling: notes as .pdf notes as .ppt
  .....Required reading: [.pdf]
  Much more on modeling the covariance structure of images: notes as .pdf
  .....Required reading: [.pdf]
  

• March 16
  Lecture 9: Non-linear Dimensionality Reduction
  notes as .ppt , notes as .pdf
  Required Reading: ISOMAP [article on web]
  Required Reading: Local Linear Embedding [.pdf]
  Required Reading: Stochastic Neighbor Embedding [.ps]
  Required Reading: Visualizing Data using t-SNE [.pdf]
  Optional reading: A paper that relates SNE to Laplacian Eigenmaps [.pdf]
  Optional reading: Dimension Reduction: A Guided Tour. [.pdf ]
  

• March 22
  Lecture 10: Collaborative filtering and Missing Data Problems
  notes as .pdf
  Required reading: Learning from incomplete data. [pdf]
  Recommended reading: Collaborative filtering and the missing at random assumption. [pdf]

• March 30
  Lecture 11a: Priors and Prejudice
  notes as .ppt , notes as .pdf
  Lecture 11b: Adaptation at multiple time scales
  notes as .ppt , notes as .pdf
  

• April 6
  Final Test (1.10pm-2.10pm in the same place as the lectures)
  Lecture 12 (2.20pm-3.00pm): The right way to recognize shapes
  

• Friday April 20: Project due at Pratt 290G by 5.00pm