behold

THE REFERENT

This is the website of Cosmo Bobak, a computer person with an interest in artificial intelligence & low-level programming, particularly deep learning and chess engines.

I care very much about the correctness and performance of software, and hope that you do too. I have a continuing fondness for the Rust programming language, and use it extensively in my projects. I work on web services, command-line tools, accidentally standardise data formats, and reïmplement the occasional ML paper.

PRINCIPAL ENDEAVOUR

VIRIDITHAS
A highly competitive chess-playing program that participates in top chess engine competitions.

an artistic image of the weights of a neuron in Viridithas

RECENT MISADVENTURES

Solving the problem of induction
Goodman’s grue and Solomonoff’s spinning razor factory
Target weighting in NNUE
Does it matter whether you predict the true future, or a stronger version of yourself?
Network architecture: multilayer
Deeper neural networks for chess engines.
ANNUEP: New main net, scaling, community results
King buckets, output buckets, and Elo scaling with network size.
ANNUEP: initial results
Results from activation functions, learning rates, and batch sizes.
The advanced NNUE program
A statement of general intent.
NNUE performance improvements
How to make NNs run fast in chess engines.
Contra “Grandmaster-Level Chess Without Search”
A critique of DeepMind's searchless chess paper, with comparison to Lc0.
An overview of computer chess techniques
Alpha-beta, PVS, iterative deepening, move ordering, search reductions, etc.
and more at the SITEMAP.