Machine Learning 2

6 EC

Semester 1, period 1

52042MAL6Y

Owner Master Artificial Intelligence
Coordinator dr. J.M. Mooij
Part of Master Artificial Intelligence,

Course manual 2019/2020

Course content

This course continues where Machine Learning 1 stopped. We will treat chapters 2,8,9,10,11,13 of the book 'Pattern Recognition and Machine Learning' by C. Bishop, plus some additional material to be distributed during class. Topics include:

  • Common probability distributions
  • Independent components analysis
  • Graphical models: Bayesian networks and Markov networks
  • Latent variable models
  • Expectation maximization
  • Variable elimination, belief propagation
  • Approximate variational inference
  • Sampling methods, MCMC
  • Sequential models: HMM and Kalman filters
  • Causality

Study materials

Literature

  • C.M. Bishop, 'Pattern Recognition and Machine Learning', 2006, Springer, ISBN 0-38-731073-8

Syllabus

Objectives

  • Gain an advanced level of understanding of the principles of machine learning;
  • Acquire the skills to apply machine learning to complex real world problems.
  • Use advanced machine learning techniques to analyze complex data and evaluate the resulting models.

Teaching methods

  • Lecture
  • Computer lab session/practical training
  • Self-study
  • Seminar

Lectures provide an overview of the material, reading the book helps understanding the details, homework sessions let you test and develop your understanding of the theory and lab sessions develop your implementation skills.

Learning activities

Activity

Hours

Deeltoets

0

Hoorcollege

28

Laptopcollege

14

Tentamen

3

Vragenuur

2

Werkcollege

14

Self study

107

Total

168

(6 EC x 28 uur)

Attendance

This programme does not have requirements concerning attendance (OER part B).

Assessment

Item and weight Details

Final grade

0.8 (80%)

Tentamen

Must be ≥ 5.5, Mandatory

0.1 (10%)

Homework

Bonus

0.2 (20%)

Computer labs

Homework assignments count as bonus. Each week, a random selection of homework exercises will be graded (unknown to the students in advance). The average grade obtained will be added as a bonus (max 1 pt.) to the final grade (before rounding).

Assignments

Homework assignments have to be made individually. Lab assignments can be made and handed in in pairs.

Fraud and plagiarism

The 'Regulations governing fraud and plagiarism for UvA students' applies to this course. This will be monitored carefully. Upon suspicion of fraud or plagiarism the Examinations Board of the programme will be informed. For the 'Regulations governing fraud and plagiarism for UvA students' see: www.student.uva.nl

Course structure

Weeknummer Onderwerpen Studiestof
1 information theoretic quantities, important probability distributions, exponential families, independent component analysis (ICA) 1.6 + 2.1, 2.2, 2.3, 2.4 + chapter 34 in MacKay
2 graphical models (Bayes nets, MRFs, factor graphs): definitions, (conditional) independences, (d-)separation 8.1, 8.2, 8.3, bits of 8.4
3 exact inference in graphical models (belief prop on trees, variable elimination, loopy bp) 8.4 + chapter 20 of Murphy
4 learning/approx inference: EM (general), VEM, variational Bayes, VAEs 9.3, 9.4, 10.1 + VAE notes + handout
5 sampling: standard, rejection, importance, ancestral, MCMC, Gibbs 11.1, 11.2, 11.3
6 HMMs + LDSs 13.1, 13.2, 13.3
7 Causality (Simpson's paradox, causal Bayes nets, truncation theorem, backdoor criterion) slides on causality + exercise1 + exercise2
8    

Timetable

The schedule for this course is published on DataNose.

Additional information

Prior knowledge: It is required to have successfully finished "Machine Learning 1" before one can register for "Machine Learning 2". The student should also be familiar with linear algebra, probability theory, calculus and programming in python.

Contact information

Coordinator

  • dr. J.M. Mooij