Search found 634 matches

by Desperado
Mon Dec 10, 2018 8:21 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Gradient Descent Introduction
Replies: 4
Views: 936

Re: Gradient Descent Introduction

Thanks both of you so far. Unfortunately my time is very limited, so there might be a delay for further posts. Now,first, i would like to follow Matthew's description and i would appreciate if you can guide me through. Let's start very simple, and say each training example has an input x (let's say ...
by Desperado
Sun Dec 09, 2018 2:32 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Gradient Descent Introduction
Replies: 4
Views: 936

Gradient Descent Introduction

Hello everybody, i am interested in a simple gradient descent implementation. Unfortunately i am not able to put the puzzle pieces together. Here is what i think that i understand and what i can do for now: base model: 1. i have a sample set of positions including results 2. i have a parameter list ...
by Desperado
Mon Oct 29, 2018 7:57 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: very small bitboard move/attack generator
Replies: 10
Views: 2016

Re: very small bitboard move/attack generator

// simple direct computations for sliders (unify bishop and rook for queen attacks) uint64_t Eng::gen_b_attack(int s, uint64_t o) { uint64_t tmp, att = 0; remove_bit(&o,s); tmp = Setbit64(s); while(False(tmp & (o | BB_HF | BB_R8))) att |= tmp <<= 9; tmp = Setbit64(s); while(False(tmp & (o | BB_AF | ...
by Desperado
Wed Jun 07, 2017 4:00 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Texel tuning method question
Replies: 63
Views: 14914

Re: Texel tuning method question

If i come close with my understanding which would mean a static score from the pv leaf, i bet it provides the same noise as any other static score you compute in any other situation. What you seem to miss is that the tuning method doesn't use only the score, but also the gradient of the score, i.e....
by Desperado
Wed Jun 07, 2017 3:44 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Texel tuning method question
Replies: 63
Views: 14914

Re: Texel tuning method question

Hi Alvaro,
... 'll be happy to explain technical details about how the gradient can be computed, if that might help...
Of course 8-) , thx for your patience.
by Desperado
Wed Jun 07, 2017 3:37 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Texel tuning method question
Replies: 63
Views: 14914

Re: Texel tuning method question

I understand that. But what is the advantage of doing a search, then, instead of directly calling static eval for each (root) position and calculating the gradient of the corresponding error function, based on initially gathering reference eval scores? The QS root might not be quiet. It would intro...
by Desperado
Wed Jun 07, 2017 3:10 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Texel tuning method question
Replies: 63
Views: 14914

Re: Texel tuning method question

So at first glance i don't know in what way the "eval at the end of the pv" is different to the search result score. It isn't. But after obtaining the PV in the first iteration I don't then actually do another search after adjusting the parameters. Each gradient descent step just does a static eval...
by Desperado
Wed Jun 07, 2017 2:57 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Texel tuning method question
Replies: 63
Views: 14914

Re: Texel tuning method question

I understand that. But what is the advantage of doing a search, then, instead of directly calling static eval for each (root) position and calculating the gradient of the corresponding error function, based on initially gathering reference eval scores? The QS root might not be quiet. It would intro...
by Desperado
Wed Jun 07, 2017 12:37 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Texel tuning method question
Replies: 63
Views: 14914

Re: Texel tuning method question

Hello Alvaro, hello HG, i don't want to be unfriendly now, especially because i simply do not understand what you are talking of, but it reads for me like "blablabla" :) I think the Texel method is operating on a set of positions with a corresponding score n x (p,s) , and it does not matter how the...
by Desperado
Wed Jun 07, 2017 11:59 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Texel tuning method question
Replies: 63
Views: 14914

Re: Texel tuning method question

Hello Alvaro, hello HG, i don't want to be unfriendly now, especially because i simply do not understand what you are talking of, but it reads for me like "blablabla" :) I think the Texel method is operating on a set of positions with a corresponding score n x (p,s) , and it does not matter how the ...