The amount of consciousness that a neural network S has is given by phi=MI(A^H_max;B)+MI(A;B^H_max), where {A,B} is the bipartition of S which minimises the right hand side, A^H_max is what A would be if all its inputs were replaced with maximum-entropy noise generators and MI(A,B)=H(A)+H(B)-H(AB) is the mutual information between A and B and H(A) is the entropy of A. 99.9%
The amount of consciousness that a neural network S has is given by phi=MI(A^H_max;B)+MI(A;B^H_max), where {A,B} is the bipartition of S which minimises the right hand side, A^H_max is what A would be if all its inputs were replaced with maximum-entropy noise generators and MI(A,B)=H(A)+H(B)-H(AB) is the mutual information between A and B and H(A) is the entropy of A. 99.9%