Search found 25 matches

by jackd
Sun Feb 03, 2019 3:20 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: So, how many of you are working on neural networks for chess?
Replies: 22
Views: 5191

Re: So, how many of you are working on neural networks for chess?

Could you summarize how Leela trains off pgns? In terms of the actual net training process, Leela learns pretty much the same way with PGNs. The major practical difference is that the PGN game positions must first be converted into the input bit planes which are fed to the Leela net training code (...
by jackd
Sat Feb 02, 2019 10:19 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: So, how many of you are working on neural networks for chess?
Replies: 22
Views: 5191

Re: So, how many of you are working on neural networks for chess?

Its not the NN evaluation that is the problem. Its the monte-carlo tree search, and it cannot be solved with tweaks to the evaluation or the monte-carlo tree search. The solution is to stop using MCTS. It is a provably inferior tree search algorithm. The fact that its well suited to parallelization...
by jackd
Sat Feb 02, 2019 10:01 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: So, how many of you are working on neural networks for chess?
Replies: 22
Views: 5191

Re: So, how many of you are working on neural networks for chess?

I think still NN engines even on higher-end commercial graphics cards play blunders at a pretty high rate. Its not the NN evaluation that is the problem. Its the monte-carlo tree search, and it cannot be solved with tweaks to the evaluation or the monte-carlo tree search. The solution is to stop us...
by jackd
Sat Feb 02, 2019 7:28 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: So, how many of you are working on neural networks for chess?
Replies: 22
Views: 5191

Re: So, how many of you are working on neural networks for chess?

I think still NN engines even on higher-end commercial graphics cards play blunders at a pretty high rate. Its not the NN evaluation that is the problem. Its the monte-carlo tree search, and it cannot be solved with tweaks to the evaluation or the monte-carlo tree search. The solution is to stop us...
by jackd
Sat Feb 02, 2019 6:52 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: So, how many of you are working on neural networks for chess?
Replies: 22
Views: 5191

Re: So, how many of you are working on neural networks for chess?

One of the main reasons I am not working on an NN based engine is that I feel I don't have the resources. I think very few people have the resources to seriously train and test a NN based engine. Reinforcement learning based NN engines are even worse in that regard. I don't understand how it matter...
by jackd
Sat Jan 05, 2019 11:47 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: mcts question
Replies: 2
Views: 900

Re: mcts question

A: After tree policy, I am at a leaf node (for my tree in memory) and I want to expand. I figured a good idea would be expanding all children in parallel, but when you expand more than one child visit counts become an inaccurate way to judge the exploration of a given node because leaves have diffe...
by jackd
Sat Jan 05, 2019 3:13 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: mcts question
Replies: 2
Views: 900

mcts question

I recently tried to add mcts to my program. My eval sucks, so the alpha-beta searches I do during expansion suck, so I don't know if my mcts algorithm works or not. Anyway here are two subtleties I need to deal with: A: After tree policy, I am at a leaf node (for my tree in memory) and I want to exp...
by jackd
Sat Dec 29, 2018 3:58 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: neural network architecture
Replies: 7
Views: 1830

Re: neural network architecture

This question, and my previous, were more suited for stack overflow, but I am going to post again here to clear up my last post . Consider the following pseudocode for standard backprop layer1 = input netlayer2 = layer1 * matrix1 layer2 = act( netlayer2 ) netlayer3 = layer2 * matrix2 layer3 = act( n...
by jackd
Fri Dec 28, 2018 4:27 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: neural network architecture
Replies: 7
Views: 1830

Re: neural network architecture

I recently had stockfish label 10 million positions. Stockfish was told "go depth 10" and then I used the depth 10 centi-pawn score as the value of this position. I wanted to keep the network input simple and didn't included castling rights or turns as input. So the castling rights were invalidated ...