Machine learning / 2026 / Runs natively
Chess ML
A neural network that learns chess by distilling Stockfish, with MCTS on top and a proper Elo measurement against Stockfish itself.
01
The problem
Hobby chess engines report accuracy on held-out data, which says nothing about playing strength.
02
How I approached it
Distil Stockfish evaluations into a network, add MCTS at inference, then measure the result the only way that counts: matches against Stockfish at known levels, converted to an Elo estimate.
03
The outcome
A strength number that comes from games played, not validation loss.
Related work
An image classification pipeline trained to recognise meme formats from a curated dataset.
A reverse GitHub for AI-written work. Give it a repository, gist, pull request, package or web page and it returns a calibrated estimate of whether the work was AI-generated, with the file, commit or line that shows why.
A 26 KB neural network running entirely in a browser tab. Draw a digit, it classifies on device, and the page reports which execution provider it actually got and what the inference cost.