This paper proposes an approach to construct a better Semantic Perceptron Net (SPN) used for topic spotting. To accomplish this task a learning paradigm call: neural network ensembling is used. Applying this technique to the original structure of Semantic Perceptron Net a new system called GA-SPN (Genetic Algorithm based Semantic Perceptron Net) was developed. The new system uses a neural netwo...

Neural networks have recently re-emerged as a powerful hypothesis class, yielding impressive classification accuracy in multiple domains. However, their training is a non-convex optimization problem which poses theoretical and practical challenges. Here we address this difficulty by turning to “improper” learning of neural nets. In other words, we learn a classifier that is not a neural net but...

This paper describes the implementation of a fast neural net simulator on a novel parallel distributed-memory computer. A 60-processor system, named MUSIC, 1 is operational and runs the back-propagation algorithm at a speed of 247 million connection updates per second (continuous weight update) using 32 bit oating-point precision. This is equal to 1 GGops sustained performance. The complete sys...

In this paper we address the problem of constructing reliable neural-net implementations, given the assumption that any particular implementation will not be totally correct. The approach taken in this paper is to organize the inevitable errors so as to minimize their impact in the context of a multiversion system, i.e., the system functionality is reproduced in multiple versions, which togethe...

2 Neural Nets, Statistical Physics and Optimization It has recently been shown that neural net models can be based upon the principles of statistical physics, In particular the spin sociative memories and optimization networks. The simulation of such neural networks promises to provide valuable insights into human and machine vision. However, for a large amay, the calculations involved become p...

Previous work on nets with continuous-valued inputs led to generative procedures to construct convex decision regions with two-layer perceptrons (one hidden layer) and arbitrary decision regions with three-layer perceptrons (two hidden layers). Here we demonstrate that two-layer perceptron classifiers trained with back propagation can form both convex and disjoint decision regions. Such classif...

In this paper a discrete competitive neural network is introduced to calculate the optimal robot arm movements for processing a considered commitment of tasks, using the branch and bound methodology. A special method based on the branch and bound methodology, modified with a travelling path for adapting in the neural network, is introduced. The main neural network of the system consists of diff...