Posts

Essentials of signal-plus-noise models: best linear unbiased predictors, innovations, filtering, prediction, smoothing, and fixed-lag smoothing.

A unifying framework and literature review for methods that perform Bayesian online learning in non-stationary environments.

A detailed and simplified introduction to the Bayesian online changepoint detection (BOCD) model for binary data using a Beta-Bernoulli model.

An exponentially-weighted moving average (EWMA) is a special case of the Kalman filter (KF). The weighted-observation likelihood filter (WoLF) is an outlier-robust variant of the KF. Here, we create a 1D version of WoLF that resembles the EWMA and is robust to outliers.

Derivation of the Kalman filter algorithm under a linear state-space model with Gaussian priors.

Introduction to the Einstein summation (einsums) and their use in machine learning.

Personal logs while reading the book Pattern Recognition and Machine Learning (PRML) by Christopher Bishop.