6.891 Machine Learning

Classes

Lectures will be held every Tuesday and Thursday 2:30pm-4pm in room 4-149.

Recitations will be held every Wednesday and Friday 1pm-2pm in room 34-302. Recitations will generally be a review of background material and material presented in lecture. Students are encouraged to attend one recitation each week.

Lecture Notes

  • Lecture 1 - 9/7/00
  • Lecture 2 - 9/12/00 (Linear Regression, Additive Models, ML Estimation)
  • Lecture 3 - 9/14/00 (Classification, Logistic Regression)
  • Lecture 4 - 9/19/00 (Neural Networks, Regularization)
  • Lecture 5 - 9/21/00 (Model Estimation)
  • Lecture 6 - 9/26/00 (Feature Selection, Bayesian Estimation)
  • Lecture 7 - 9/28/00 (Guest lecture: Prof. Paul Viola)
  • Lecture 8 - 10/3/00 (Chapters 5.1--5.5 Margin etc., 5.11 SVMs, 9.5 Boosting)
  • Lecture 9 - 10/5/00 (Chapter 5.11 SVMs; Pages 1-8 of C. Burges tutorial: ps-file, gzipped ps-file)
  • Lecture 10 - 10/12/00 (VC-Dimension, Model Selection; Chapters 3.1-3.5, 4.3)
  • Lecture 11 - 10/17/00 (Density Estimation, Expectation-Maximization)
  • Lecture 12 - 10/19/00 (Clustering)
  • Lecture 13 - 10/24/00 (Clustering, hierarchical mixture models, mixtures of experts)
  • Lecture 14 - 10/31/00 (Markov chains, HMMs; handout paper)
  • Lecture 15 - 11/2/00 (HMMs cont'd, biosequence models)
  • Lecture 16 - 11/7/00 (dynamic programming, biosequence HMM, representation, graphical models)
  • Lecture 17 - 11/9/00 (graph models, bayesian networks)
  • Lecture 18 - 11/14/00 (medical diagnosis, markov random fields)
  • Lecture 19 - 11/16/00 (markov random fields; image reconstruction)
  • Lecture 20 - 11/21/00 (Junction Trees; inference and estimation in graphical models)
  • Lecture 21 - 11/30/00 (Guest lecture: Prof. Leslie Kaelbling)
  • Lecture 22 - 12/5/00 (estimation and model selection in graph models)
  • Final exam - 12/7/00
  • Lecture 23 - 12/12/00 (research problems in machine learning; graded final exam, problem set #4)
  • Handouts
  • Administrivia
  • See the coursework section for homework handouts.