Getting Started

Greetings, internet denizens. I have intended to put up a public web presence for some time for a variety of reasons. I am intending for this space to be a sort of project log for the things I want to make, whether they be hardware or software, music, art, and maybe some attempts at useful technology.

The current motivating factor that has spurred me on is the fact that I hope this log of project notes will encourage me to be more focused in my hobby efforts. I'm also a believer in the principles of open source hardware and software, and I would like things that I learn to be useful to other people.

I'm going to kick things off witha series of posts titled: learning machine learning with entirely too many MNIST digits, which is a nod to the great Learning Rust With Entirely Too Many Linked Lists. This series of posts will be my opinionated (read: naive and ill-informed) take on some of the current ML frameworks, and we'll be covering the following topics:

• Convolutional Neural Nets
• Multilayer Perceptrons
• Probabilistic Graphical Models
• etc.

with the following frameworks:

• PyTorch
• TensorFlow
• and other esoteric things like Juice

The main test problem will be MNIST digit classification, because of course it will.

These posts won't necessarily be coming out in any set timeframe or order, and I reserve the right to talk about anything I want in between posts in the series.

TTFN.