It was suggested to me recently that to fit a predictive model supposes that the domain variables cause the image variables. This is not the case. Predictive models assume nothing about causality ...

I tried googling “what is a picture made with crayon called” and I consistently got suggestions for crayon etching, which is a more specific usage of crayons. I asked ChatGPT the following question...

It is well-known that neural networks are universal function approximators under mild assumptions. Such theorems proving arbitrary bounds on the empirical risk and similar results tell us that a go...

Introduction This post is about comparing and contrasting my view of what a statistical model is with how Dustin Fife views statistical models. One of the keystones of Dustin’s perspective is that...

Some of the figures in this post are more easily read in light mode rather than dark mode. When time allows I will find suitable replacement images. Back in May 2020 I released a video on YouT...

I finished my 7th painting! There are certaintly a lot of flaws in this painting, but I am glad I got to the end of it. A few things went wrong on me. If you look carefully at the trees in the ...

Nothing like an assortment of miscellaneous quotes, right? I don’t entirelly agree with them, nor with everything else the people who said them have said, but these quotes have made me think or fee...

Suppose we have a collection of IID random variables ${ X_1, \ldots, X_n }$, and we also have a second collection of IID random variables ${ Y_1, \ldots, Y_m }$. Each $X_i \sim F_X$ and $Y_i \sim F...

This post just shows what a Dirac delta distribution is and how to make a mixture of Dirac delta distributions in PyMC. The Dirac delta distribution is somewhat unlike most distributions you will ...

Introduction This post looks at the notion of ranking data or variables. Broadly, ranks are numbers used to describe something about the order-like properties on the elements of a set. We might th...