Welcome to my blog!
Where I mostly write about deep learning, math, and physics.
Posts
Beyond Loss Minimization: A Bayesian View of Deep Learning
An introduction to Bayesian concepts in deep learning that links marginal likelihood, curvature, and effective dimensionality to the geometry of flat minima and generalization.
Read more →
Image by Google DeepMind on Unsplash
First-Principles Analysis of an Idealized BLDC Motor
From Maxwell's equations to torque and back-EMF, and the relationship between motor constants.
Read more →
Photo by Mika Baumeister on Unsplash
A Method for Text to Text Encryption Using LLMs
In this article, I explore a fun and probably useless application of language models: encrypting messages in natural language. This article includes a demo with an encrypted chat.
Read more →
Why is the Adam Optimizer Working so Well?
In this article I present a Bayesian derivation of the Adam optimizer. I think is a more convincing explanation of it's success compared to more traditional derivations which use a bound on the regret.
Read more →
A Bayesian Derivation of the Kalman Filter
A hopefully intuitive derivation of the Kalman filter equations from Bayesian principles.
Read more →
©ESA/Rosetta/NAVCAM
A Simple Derivation of the Woodbury Matrix Identity
A simple and hopefully intuitive derivation of the Woodbury matrix identity.
Read more →
Tools
LaTeX to SVG
LaTeX to SVG path converter
Try it →
