**MOVED TO https://github.com/LeelaChessZero/leela-chess ** A chess adaption of GCP's Leela Zero
Role in this project:
Back-end Developer Contributions:2 releases, 96 commits, 139 PRs in 3 months
Contributions summary:Andy primarily contributed to the `leela-chess` project by modifying and improving core chess engine functionalities. Their work included normalizing policy outputs for new randomized neural networks and enhancing the `dump_stats` feature by adding increased move statistics and fixing display issues. Further contributions involved optimizing training output, adding and integrating the `NNCache` for performance improvements, and fixing UCI command outputs, demonstrating a focus on the engine's core mechanics, performance, and user interface.
chess-enginegcpzerouci
Go engine with no human-provided knowledge, modeled after the AlphaGo Zero paper.
Role in this project:
Back-end Developer Contributions:20 commits, 23 PRs, 555 comments in 3 months
Contributions summary:Andy primarily contributed to the core logic of the Go engine, focusing on refining the resign behavior and adding supporting tools for analysis. They implemented changes to the resign conditions within the UCTSearch algorithm, altering the criteria for game resignation based on the winrate. Additionally, the user added a Python script, resign_analysis.py, designed for analyzing data related to resignation thresholds, demonstrating a focus on analyzing and optimizing the engine's decision-making processes. Furthermore, they introduced changes in autogtp/main.cpp, src/Training.cpp, src/GTP.cpp, src/UCTNode.cpp and other files, which indicates a deep understanding of the engine's internal workings and testing/debugging.
golangalphago-zerozeroalphagoknowledge