Search found 206 matches

by Tony P.
Thu Oct 08, 2020 7:10 am
Forum: Computer Chess Club: General Topics
Topic: Which of the many chess engines in this forum use b strategy ?
Replies: 92
Views: 8380

Re: Which of the many chess engines in this forum use b strategy ?

Todays's chess engines make breathtaking, beautiful moves. I have seen Komodo make a queen sacrifice that stunned the crowd. There are Morphy like tactical stunners. There are Botvinick like positional squeezes. There are strategic masterpieces and even utter nonsense gaffes. I think you have not b...
by Tony P.
Wed Oct 07, 2020 10:10 pm
Forum: Computer Chess Club: General Topics
Topic: Question about nnue
Replies: 6
Views: 841

Re: Question about nnue

We'll see. I do think Arasan needs a neural upgrade to stay competent; fortunately, its elaborate position scoring routines are a natural fit for the preprocessing of positions before they're put into an NN. If I'm not mistaken, the rest of the active top-20 engines (except Booot that's hopeless bec...
by Tony P.
Wed Oct 07, 2020 9:11 pm
Forum: Computer Chess Club: General Topics
Topic: A Crossroad in Computer Chess; Or Desperate Flailing for Relevance
Replies: 152
Views: 21157

Re: A Crossroad in Computer Chess; Or Desperate Flailing for Relevance

Actually, one of the advantages of the ReLU activation is that it maps all the negative inputs to the same zero output, so if an input was negative and remains negative after an incremental update, then the output remains zero, and fewer recalculations are needed in further layers.
by Tony P.
Wed Oct 07, 2020 9:01 pm
Forum: Computer Chess Club: General Topics
Topic: A Crossroad in Computer Chess; Or Desperate Flailing for Relevance
Replies: 152
Views: 21157

Re: A Crossroad in Computer Chess; Or Desperate Flailing for Relevance

Thanks for correcting my blunder: from the perspective of programming, it's indeed better to view biases as a constant vector added to the matrix-vector product, as then one can make use of SIMD if the activation function is the same for all the components (or if there are only a few different funct...
by Tony P.
Wed Oct 07, 2020 8:40 pm
Forum: Computer Chess Club: General Topics
Topic: A Crossroad in Computer Chess; Or Desperate Flailing for Relevance
Replies: 152
Views: 21157

Re: A Crossroad in Computer Chess; Or Desperate Flailing for Relevance

A neuron in a net is a unit with some inputs that can be activated or not (0, 1) and an output that is also 0 or 1. That's a binary neuron, typically used on resource-constrained devices like mobile, where full-accuracy NNs consume too much memory. Most commonly used neurons use floating-point numb...
by Tony P.
Wed Oct 07, 2020 6:01 pm
Forum: Computer Chess Club: General Topics
Topic: Minic version 2
Replies: 157
Views: 45923

Re: Minic version 2

Furthermore, they say on http://rebel13.nl/misc/sim2019.html. "Note the extreme high similarity of 83.57% between Schooner and Xiphos 0.3" 83.57% is quite a number. I don't know what to think about it. Ed admitted the error in EO: Dennis, I reran the SIM-test with Schooner 2.0.34 and also included ...
by Tony P.
Tue Oct 06, 2020 8:11 pm
Forum: Computer Chess Club: General Topics
Topic: Which of the many chess engines in this forum use b strategy ?
Replies: 92
Views: 8380

Re: Which of the many chess engines in this forum use b strategy ?

It's a great idea to let the user sculpt many different styles (or start from presets resembling past human masters), thanks for the remark :)
by Tony P.
Tue Oct 06, 2020 7:40 pm
Forum: Computer Chess Club: General Topics
Topic: Which of the many chess engines in this forum use b strategy ?
Replies: 92
Views: 8380

Re: Which of the many chess engines in this forum use b strategy ?

Just out of curiosity: if an engine is released that asks you once in a while about a game or move it's just played, 'Are you entertained?', and adjusts its style parameters automatically by analyzing your feedback, will you be happy to use it? :D
by Tony P.
Tue Oct 06, 2020 6:15 am
Forum: Computer Chess Club: General Topics
Topic: A Crossroad in Computer Chess; Or Desperate Flailing for Relevance
Replies: 152
Views: 21157

Re: A Crossroad in Computer Chess; Or Desperate Flailing for Relevance

Generally, as a TCEC spectator, I'd like them to maintain a good gameplay quality-diversity compromise (which hasn't been reached in S19 because two nets mainly trained by Sergio Vieri were allowed to face each other, as well as the LS net inside Allie), which rather has to do with how much an invit...
by Tony P.
Mon Oct 05, 2020 10:26 pm
Forum: Computer Chess Club: General Topics
Topic: Which of the many chess engines in this forum use b strategy ?
Replies: 92
Views: 8380

Re: Which of the many chess engines in this forum use b strategy ?

So you have no rule proposal? How do you expect to create a program without a good idea? The recent advances in 'learning to plan' (aka 'reward shaping' in reinforcement learning) suggest that it makes sense to let an engine learn the rule in an unsupervised or self-supervised fashion, so it's okay...