Back to Talks

Systematic approach for machine learning methods design based on potential theory

Nadia Udler

Audience level: Intermediate
Topic area: Modeling


With the increase of computer power machine learning methods become a method of choice for solving many real world problems, where previously analytical approximations would be more appropriate in terms of speed. This creates a great demand for machine learning software that is easy to use. We present an approach for constructing such methods in systematic way. We demonstrate several tutorials that help to understand essential building blocks and parameters of machine learning methods.


In this presentation we briefly describe the approach for constructing machine learning methods based on potential theory and highlight essential building blocks that comprise a base of any method. We present student tutorials that help to understand and visualize these building blocks (Householder reflection, operator of space dilation, computing natural gradient, change of system of coordinates or scaling) and provide examples of popular heuristic methods where these building blocks can be seen (Covariance Matrix Adaptation Evolution Strategy – space transformation, Nelder and Mead – reflection, Natural Gradient Evolution Strategy,etc).