2 papers
cond-mat1995
UNREALIZABLE LEARNING IN FEEDFORWARD NEURAL NETWORKS
Matts Sporre
Statistical mechanics is used to study unrealizable generalization in two large feed-forward neural networks with binary weights and output, a perceptron and a tree committee machi…
cond-mat1995
UNREALIZABLE LEARNING IN BINARY FEEDFORWARD NEURAL NETWORKS
Matts Sporre
Statistical mechanics is used to study unrealizable generalization in two large feed-forward neural networks with binary weights and output, a perceptron and a tree committee machi…