Multithreshold Neural Units and Networks

We deal with theoretical issues concerning the application of the multithreshold architecture in the theory of neural computation. The way of representing a multithreshold function by 2-layer neural network consisting of single-threshold neural units with equal weights is established in the paper. We also study the complexity of the problem of the learning k-threshold neurons and prove that this problem is NP-hard if the number of thresholds is greater than one.