Dear ASilver,
In the
response to my post to on LCZero forum, you said that's it's "not even a misunderstanding", which to my understanding means that you don't think that you did anything wrong.
In that case, to help me understand your vision of the situation, would it be possible for you to clarify your position in the following questions:
1. Do you agree that in "DeusX engine" entry submitted to TCEC, total amount of human effort inherited from LCZero and Leela Zero project, is greater than total amount of human effort that you personally contributed?
Reminder: LCZero effort includes, in particular, engine, including MCTS implementation, batching support, cudnn evaluation, movegen, uci support, optimizations and debugging; training script, both for supervised and unsupervised data; and time spent tuning and trying all that, e.g. network and training parameters, underspecified in papers from DeepMind.
2. Do you agree that given the amount of effort inherited from LCZero, and given that LCZero itself participates in TCEC too as a separate entity, it would be fair to announce which parts of your TCEC entry are a direct reuse of parts of LCZero project and how exactly they are different?
3. Why didn't you discuss possibility to use of Lc0 engine for your TCEC entry with LCZero developers? LCZero participates as a separate entity on it's own, and surely you cannot submit the engine that you didn't author without permission from the authors.
4. Why didn't you mention in the interview that you used and are using the code from LCZero project? You had lots of opportunities to do that.
5. You did not write a chess engine. Why didn't you ask TCEC admins to correct the title "
Deus X – the NN chess engine by Albert Silver" of this announcement:
http://www.chessdom.com/deus-x-the-nn-c ... rt-silver/ and a similar sentence in the article itself?
Thank you in advance for your answers.