Machine Learning

When machines learn by themselves

I put a physics-informed neural network up against a plain finite-difference solver twice: once in 1D and once in 5D. The winner changed.

A small, strange discovery in machine learning called grokking

My AI detectors flagged many genuine reviews, and filtering them made the sentiment model less accurate.

Reproducing Anthropic's "Toy Models of Superposition" from scratch in NumPy, with hand-derived gradients and no borrowed numbers.

The mechanics behind local reasoning experiments with Unsloth and why the reward function matters as much as the model.

Five scikit-learn defaults that deserve a closer look before your next model reaches production

The recurring cost of production AI is not inference. It is re-qualification: the eval reruns, prompt retuning, and regression testing you owe every time a model changes under you. Here is what that tax actually covers, and how to budget for it before it surprises you.

What ordering and not eating a large pizza tells us about ML memory management

How moving randomness outside the computation graph turns noisy gradient estimators into low-variance, differentiable ones