In cellular automata based-model: any benefit for dynamic neighborhood?

As we strive to create embodied artificial intelligence with increasingly greater capabilities, do we have a moral responsibility to set ourselves an acceptable limit to the level of intelligence we want to achieve? Is there a level of intelligence at which it becomes morally questionable to restrict a robot’s movements, to alter its program or to pull the plug? Or is it simply too early to think about such issues at all?

Given a set of SGF files for the game of GO, how should I employ this data to find the patterns, such that the computer can predict the next move when in the same situation i.e. same board state?

Firstly, l define what I mean by structure equation. The solution of an algebraic equation is a number. The solution of a functional equation is a function, for instance, differential equations. The solution of a structure equation would be a structured set ,say, a geometry, a differentiable manifold, an algebra etc.

As far as I know, the only structure equation that has been widely investigated is the Einstein one, the solution of which is a differentiable manifold determined by its connection coefficients.

A learning system, modifies its own structure as he is learning new topics and handling new information. Thus, the unknown of a learning system theory is its structure sequence determined by the acquired information. I think that this topic must be investigated.

Seeking references of sample codes as examples.

My data is very close to random and with sign function I can not have good results. I think I've lost some information that find NN with sign function.

I want to combine semantic web service with genetic algorithm for optimization of problems

There are source codes given in two text files. Is there a way to find if the logic applied in both the files is same?