[OC] Ich habe den „psychischen Stress“ eines Schachspielers aufgezeichnet, indem ich die menschlichen Vorhersagen eines neuronalen Netzwerks mit der absoluten Wahrheit eines Supercomputers verglichen habe.
Von lpshred
[OC] Ich habe den „psychischen Stress“ eines Schachspielers aufgezeichnet, indem ich die menschlichen Vorhersagen eines neuronalen Netzwerks mit der absoluten Wahrheit eines Supercomputers verglichen habe.
Von lpshred
3 Kommentare
Source: I designed the metrics and built this custom pipeline using Python (with AI assistance for the syntax and deep math), Pandas, and Plotly/Dash. It compares the Stockfish 16.1 engine against the Maia Neural Network to measure human error and stress.
If anyone wants to see the code, I put the GitHub link in my Reddit profile bio!
I don’t know enough about chess to know if the data is correct, but why is black under so much more pressure in the entire game? Like, sure, at the start of the game, you are a move behind. That makes sense. But after 10 or 20 rounds, that 1 move should mean less, but it doesn’t
I don’t understand what the title is saying. Are you plotting a neural network’s *prediction* of how stressed a player would be based on the game state? Or is this actual data using biological measurements?