Re: The status of gnubg?

[Aaron Tikuisis]

>From my understanding based on Joseph Heled's page, the NN training does not try to make the %s match the true values, but rather tries to ensure that the NN makes the best decisions. (This is a subtle difference, and it doesn't seem to fundamentally change the question.)

I imagine that as the NN improves, so does the benchmark.

How often/recently is the gnubg NN retrained? Do we think it can be improved much?

Best, Aaron

---

From: Joaquín Koifman <pamalejo@gmail.com>
Sent: October 16, 2020 12:12 PM
To: Ian Shaw <Ian.Shaw@riverauto.co.uk>
Cc: Aaron Tikuisis <Aaron.Tikuisis@uottawa.ca>; Øystein Schønning-Johansen <oysteijo@gmail.com>; bug-gnubg@gnu.org <bug-gnubg@gnu.org>
Subject: Re: The status of gnubg?

 

Attention : courriel externe | external email

May I ask a couple of questions regarding NN training?

From my little understanding, I suppose there are 2 sources of errors during an evaluation:

a) the NN may be intrinsically unable (say, because of the type, number, etc of inputs) to "score" well, when compared with the true equity/%s (at least, the ones you want to replicate) of a position, or

b) the benchmark against which the NN is being tested might not have the "true" equity/%s because, for example, the rollouts were done in 0-ply.

Is there any way to know which of these two factors is limiting the most the improvement of gnubg's NN? I mean, do we need to improve the benchmark, do more training or change the NN altogether to improve the evaluation?

Thanks