10-601 Sequences

From Cohen Courses
Jump to navigationJump to search

This a lecture used in the Syllabus for Machine Learning 10-601B in Spring 2016

Slides

Optional Readings

  • This is not covered in Mitchell. For HMMS: Bishop 13.1-13.2 cover this material. The most-used introduction to HMMs is: Rabiner, Lawrence R. "A tutorial on hidden Markov models and selected applications in speech recognition." Proceedings of the IEEE 77.2 (1989): 257-286. There is a nice general introduction to CRFs by Sutton and McCallum.

Summary

You should know:

  • The definition of an HMM
  • What the Viterbi and forward-background algorithms are:
    • What their complexity is
    • What they compute.
  • How to learn the parameters HMMs when
    • The states associated with the training data are observed
    • The states are unobserved.
  • What the advantages of a CRF are compared to an HMM.
  • How HMMs and CRFs relate to naive Bayes, logistic regression, and generative and discriminative models.
  • How HMMs or CRFs can be used for named-entity recognition (NER) and other sequential classification tasks.