In 2018, Yu Nasu proposed a neural network based evaluation function for computer shogi in his paper "NNUE: Efficiently Updatable Neural-Network-based Evaluation Function for computer Shogi".

This appraoch proved to be very powerful, and has been backported to the chess program Stockfish.

The original paper is available in Japanese.

This repo contains translations of the Japanese paper to English as well the TeX source file.

Links to the PDF:

• English: NNUE paper

2025-09-19: Finally found time to properly typeset the translation in TeX. Removed the German document.