Institutionernas doktorandkurser

Startdatum och periodicitet för kurser kan variera. Se detaljer för respektive kurs för aktuell information. För anmälan, kontakta respektive kursansvarig.


Pattern Classification and Machine Learning

  • Kurskod: FSSY070
  • ECTS-poäng: 8,0
  • Institution: ELEKTROTEKNIK
  • Forskarskola: Signaler och system
  • Periodicitet: The course will be given in the spring 2018.
  • Undervisningsspråk: Kursen kommer att ges på engelska
Course content:
1. Introduction: outline, "road map" for classification methods
 Signal characterization, discriminative features, feature normalization, examples.
2. Supervised learning methods
 ML, Bayesian parameter estimation for pattern classification;
 Statistical learning and boosting methods (Support Vector Machines and Adaboost);
 Bayesian and belief networks;
 GMMs and EM algorithm.
3. Unsupervised learning methods
 ML parameter estimation for component densities from the mixture density;
 Clustering methods (K-means, mean-shift);
 Topic models and BoW (bag of words);
4. Automated learning of features
 NNs and deep learning
Litteratur

Course book/material:
1. Richard O. Duda, Peter E. Hart, David G. Stork,  Pattern classification, John Wiley & Sons, 2nd Ed.
2. Christopher M. Bishop, Pattern Recognition and Machine Learning, Springer.
3. David Barber, Bayesian reasoning and machine learning, Cambridge university press, 2012.
4. Related state-of-the-art publications and tutorial materials
5. Matlab toolbox for PR.


Home exercises and Exam:
 Solving several theoretical problems
 Matlab programs on several selected topics and presentations
 Written exam: assignment of problem solving and programming tasks.

Föreläsare
Irene Gu
Mer information
Irene Gu
E-mail: irenegu@chalmers.se

Publicerad: to 14 okt 2010. Ändrad: on 23 aug 2017