flâneur — a map of the web's best reading

Lewis Smith - A gentle introduction to information geometry

robots.ox.ac.uk · 6,159 words · saved by 1 readers

The following post is a slightly modified version of some notes I made for an internal reading group at Oxford. I actually gave the tutorial ages ago, but I finally got around to cleaning it up and putting it online. In brief, information geometry is the study of the geometrical structure of families of probability distributions. The field thus brings together statistics, information theory and differential geometry, revealing some fascinating and unexpected connections between them. This tutorial attempts to fit a flavour of a fairly technical and abstract field into a brief tutorial that assumes no knowledge of differential geometry. Therefore, I have chosen intuition over rigour whenever I felt it appropriate, for which I make no apologies, and only introduced the absolute minimum of background. I skip over a lot of very interesting things to try to get to the information metric as quickly as possible. There is a list of further reading in the conclusion for those who are interested

Posted on September 27, 2019 The following post is a slightly modified version of some notes I made for an internal reading group at Oxford. I actually gave the tutorial ages ago, but I finally got around to cleaning it up and putting it online. Introduction In brief, information geometry is the study of the geometrical structure of families of probability distributions. The field thus brings together statistics, information theory and differential geometry, revealing some fascinating and unexpected connections between them. This tutorial attempts to fit a flavour of a fairly technical and abs

Explore this link on the map →

saved by

related reading