CloseHelpPrint
Kies de Nederlandse taal
Course module: 2DI70
2DI70
Statistical learning theory
Course info
Course module2DI70
Credits (ECTS)5
Category-
Course typeGraduate School
Language of instructionEnglish
Offered byEindhoven University of Technology; Mathematics and Computer Science; Mathematics;
Is part of
Computer Science and Engineering
Data Science in Engineering (CSE)
Industrial and Applied Mathematics
Contact persondr. R.M. Pires da Silva Castro
Telephone2499
E-mailr.m.pires.da.silva.castro@tue.nl
Lecturer(s)
Responsible lecturer
dr. R.M. Pires da Silva Castro
Feedback and reachability
Other course modules lecturer
Contactperson for the course
dr. R.M. Pires da Silva Castro
Other course modules lecturer
Academic year2016
Period
3  (06/02/2017 to 23/04/2017)
Starting block
3
TimeslotC1: C1 - Tu 1-2, Fr 5-6
Course mode
Fulltime
Remarks-
Registration openfrom 15/06/2016 up to and including 15/01/2017
Application procedureYou apply via OSIRIS Student
Explanation-
Registration using OSIRISYes
Registration open for students from other department(s)Yes
Pre-registrationNo
Waiting listNo
Number of insufficient tests-
Number of groups of preference1
Learning objectives
  • Get acquainted with the probabilistic framework of learning theory
  • Use the tools presented in class to study the performance of sev-eral machine learning algorithms
  • Understand the trade-off between approximation and estimation error (or bias and variance)
  • Read recent research articles in the area of learning theory, and be able to effectively present them in class
  • Content
  • A probabilistic view of regression and classification
  • Introduction to complexity regularization
  • Denoising smooth functions
  • Analysis of the histogram classifier
  • PAC bounds and concentration of measure
  • Analysis of Decision Trees
  • Maximum likelihood estimation and complexity regularization
  • Denoising Smooth Functions with Unknown Smoothness
  • Denoising in piecewise smooth function spaces
  • Introduction and applications of Vapnik-Chervonenkis (VC) Theory
  • Entrance requirements
    Entrance requirements tests
    -
    Assumed previous knowledge
    Successful completion of a probability theory course (bachelor level) and reasonable level of mathematical maturity. Knowledge of statistics is not necessary (although it also doesn’t hurt).
    Previous knowledge can be gained by
    -
    Resources for self study
    -
    Short promotional description of the course
    The aim of statistical learning theory is to study, in a probabilistic frame-work, the properties of learning algorithms. The purpose is two-fold: endow existing methods with performance guarantees, and suggest novel algorithmic approaches. This course gives a thorough introduction to the theory and methods of statistical learning theory, and in particular complexity regularization. This latter is at the heart of the most successful and popular machine learning algorithms today.
    Short promotional description of the course
    The aim of statistical learning theory is to study, in a probabilistic frame-work, the properties of learning algorithms. The purpose is two-fold: endow existing methods with performance guarantees, and suggest novel algorithmic approaches. This course gives a thorough introduction to the theory and methods of statistical learning theory, and in particular complexity regularization. This latter is at the heart of the most successful and popular machine learning algorithms today.
    Bachelor College or Graduate School
    Graduate School
    Required materials
    -
    Recommended materials
    Lecture notes provided during the course
    Recommended (non-mandatory) books (available through the li-brary)
    Instructional modes
    College / course

    General
    -

    Remark
    -
    Guided selfstudy

    General
    -

    Remark
    -
    Lecture with notebook / PC

    General
    -

    Remark
    -
    Tests
    Written exam
    Test weight100
    Minimum grade6
    Test typeFinal examination
    Number of opportunities2
    OpportunitiesBlock 3, Block 4
    Test duration in minutes-

    Assessment
    -

    Remark
    in-class presentation of research articles, and a final exam. Depending on the number of students the mode of assessment

    CloseHelpPrint
    Kies de Nederlandse taal