[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

AI wins final match won series 4-1

Posted By: Frank Berger
Date: Thursday, 17 March 2016, at 9:41 p.m.

In Response To: AI wins final match won series 4-1 (Karol Szczerek)

First to chess. Chess is different to BG a highly tactical game so NN playing chess program has been pretty weak, but IIRC in the last year was one which was remarkably strong, but weaker than the traditional ones.

Now to Go: It isn't the first approach to use a NN for Go, but they failed (with the exception of Facebooks Go programm). In BG we use a simple Perceptron with one hidden layer, but alphaGo uses a deep convolutional NN with 20(? I'm to lazy to search for the exact number) layers. When I started training BG-NN I looked for alternatives for structure and training and deep nets were regarded at that time as very difficult. No the Hardware is much faster and a large cluster is easy tp get.

Messages In This Thread

 

Post Response

Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:

 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.