...
Linear factor models Linear factor models of stock returns are commonly used by practitioners and academics. The topic is well covered (see Elements of Quantitative Investing , or this post ), so I won't repeat the basics, but instead dig a little more into an interesting result that I've seen mentioned a few times but never fully explored. Pure factor portfolios as estimators Our model for stock…
One of my least favourite activities associated with programming has always been configuration management. Each tool and platform has its own config file format (JSON, YAML, custom scripting lang, ...) with a different naming convention ( ._rc, ._.conf , ._config`, ...) in a different location with weird obtuse setting names and values (only a slight exaggeration). This is a natural consequence of…
Backpropagation is the fundamental algorithm that allows us to train neural networks, and autodifferentiation is the primary tool used to do this in practice. Lets start by writing a basic autograd library, and then use it to train a basic neural network using backpropagation. Smallgrad What is autodifferentiation An autodiff library lets us compute the gradient of some derived value with respect…
I have been using Claude Code (and infrequently Codex) fairly frequently over the last few months on a variety of side projects. Some personal reflections: These tools are extremely capable. My most recent project is a fairly complex 20k LoC machine-learning focused project. I have perhaps written 0.5% of the code by hand, the rest is entirely model generated. I am happy with the quality of the…
For the last couple of years I have been using spaced repetition to revise and learn a variety of topics - linear algebra, probability, some ML concepts and a few programming libraries amongst other things. I’ve read quite a few posts and articles about the effectiveness of spaced repetition as a learning tool, but from memory none of them highlight the “ratchet” effect that in my opinion is the…
Theory We are interested in solving the following convex optimisation problem : min x f 0 ( x ) s . t . f i ( x ) ≤ 0 , i = 1 , . . , m A x = b where x is our (primal) optimisation variable, f 0 is our objective function, f i ( x ) ≤ 0 are our m inequality…