Week 4: Sequential Decision Making Under Uncertainty: Bayesian Optimisation

[jupyter][google colab][pdf slides][pdf worksheet]

Carl Henrik Ek

Abstract:

In this lecture we will see the first example of surrogate modelling. In specific we will extend the machine learning loop to also include the data aquisition. We will the formulate a sequential decision process where we aim to find the extremum of a explicitly unknown function. In specific we will introduce the concept of Bayesian optimisation which is the technique that underpins the exciting field called Auto-ML.

publications and links will appear here.