Machine Learning II 2019/20 spring

Homework points

Link to the table

Requirements to solutions

Homeworks

• 15.02 - 22.02: I. Performance measures
  • Empirical risk and its variability
  • Convergence of empirical risk to true risk
  • Bias-variance trade-off in performance estimation
  • Applications of crossvalidation algorithms
  • Confidence intervals for crossvalidation estimates
  • Different flavours of bootstrapping algorithms
  Homework:
  • You can get up to 15 points from the homework exercises.
  • Nominal score for the homework is 10 points.
  • There are no restrictions how you choose exercises.

• 29.02 - 13.03: II. Basics of probabilistic modelling
  • True meaning of confidence intervals
  • Quantile-Quantile plot and its confidence envelope
  • Probability distributions over real-valued functions
    • Confidence envelopes for predictors
    • Confidence envelopes for ROC curves
  • Bayesian inference and its internal consistency
  • Naive-Bayes classification
  • Markov chains and detection of abnormal words

• 15.03 - 22.03: III. Sequence models and belief propagation
  • Markov chains with discrete state space
  • Hidden Markov models with discrete state space and discrete observations
  • Hidden Markov models with discrete state space and continuous observations
  • General framework for belief propagation: prior, likelihood and marginal posterior
  • Higher-order hidden Markov models and their applications.

• 25.03 - 16.04: IV. Direct applications of normal distributions

• 25.03 - 16.04: V. Normal distributions and affine projections

5. Affine data projections

• 24.04 - 03.05: VI. Model-based clustering

6. Hard clustering

• 05.05 - 24.05: VII. Expectation-Maximisation algorithm

7. Soft clustering and EM-algorithm

• 05.05 - 24.05: VIII. Expectation-Maximisation algorithm and sequential data

8. EM algorithm and sequential data