Max Welling's Classnotes in Machine Learning

Statistical Estimation [ps]
- bayesian estimation
- maximum a posteriori (MAP) estimation
- maximum likelihood (ML) estimation
- Bias/Variance tradeoff & minimum description length (MDL)

Expectation Maximization (EM) Algorithm [ps]
- detailed derivation plus some examples

Supervised Learning (Function Approximation) [ps]
- mixture of experts (MoE)
- cluster weighted modeling (CWM)

Clustering [ps]
- mixture of gaussians (MoG)
- vector quantization (VQ) with k-means.

Linear Models [ps]
- factor analysis (FA)
- probabilistic principal component analysis (PPCA)
- principal component analysis (PCA)

Independent Component Analysis (ICA) [ps]
- noiseless ICA
- noisy ICA
- variational ICA

Mixture of Factor Analysers (MoFA) [ps]
- derivation of learning algorithm

Hidden Markov Models (HMM) [ps]
- viterbi decoding algorithm
- Baum-Welch learning algorithm

Kalman Filters (KF) [ps]
- kalman filter algorithm (very detailed derivation)
- kalman smoother algorithm (very detailed derivation)

Approximate Inference Algorithms [ps]
- variational EM
- laplace approximation
- importance sampling
- rejection sampling
- markov chain monte carlo (MCMC) sampling
- gibbs sampling
- hybrid monte carlo sampling (HMC)

Belief Propagation (BP) [ps]
- converting directed acyclic graphical models (DAG) into junction trees (JT)
- Shafer-Shenoy belief propagation on junction trees
- some examples

Boltzmann Machine (BM) [ps]
- derivation of learning algorithm

Generative Topographic Mapping (GTM) [ps]
- derivation of learning algorithm

 

 

---

[Back to Max Welling's's home page]