Study Group for: Deep Learning and the Game of Go

Hi, is anyone interested in a study group for the book: Deep Learning and the Game of Go ?

I am finding the book easy, but the python is difficult for me as I am primarily a Java and C programmer.

How about you?

There is a Facebook group, but not a lot of activity there.


Well here’s a simple solution: If the concepts are easy for you, just write the thing in C.

You should, of course, write it in go.


I guess you might already know, but there is a popular C++ reimplementation of the AlphaGo Zero bot concept, called Leela Zero:

I don’t know anything about the book that you mentioned, but if you are interested in studying C++ code rather than Python examples, I would recommend checking out Leela Zero.

Also, if you are interested in learning about general “deep learning” concepts, one of the most highly-regarded books on those topics is available for free online:

I have used leela zero. I have the book.


some of the algorithms are complicated, so i want to use a library.


unfortunately python has the best interface for keras and tensorflow. the java interface is poor. dl4j is not as popular and a pain to build.


If you’re experienced with programming already, then switching to python shouldn’t be a problem, right? It’s not exactly a difficult language to learn. It’s very readable and there are no problems with anything you don’t want to spend time on while learning deep learning (e.g. memory allocation).

Especially when learning, it’s more useful to play around with the concepts than to build everything in a language of your choice.

currently trying to fix this.

going back to the latest anaconda seems to work after fixing the numpy problem.


D:\ray\dev\conradapps\dlgopy\dlatgog\code>conda install tensorflow
Solving environment: failed

UnsatisfiableError: The following specifications were found to be in conflict:

  • anaconda==2018.12=py37_0 -> bleach==3.0.2=py37_0
  • anaconda==2018.12=py37_0 -> numexpr==2.6.8=py37hdce8814_0
  • anaconda==2018.12=py37_0 -> scikit-learn==0.20.1=py37h343c172_0
  • tensorflow

back on track after a few diversions:

I hope you can get it working… If it works python does really make life easier, but I didn’t expect it to give this many problems.

On a different note, I would really recommend the following video series to anyone who is learning machine learning:

1 Like

on another note:

the download has a bot you can play as well as the source.