Loading...
Please wait, while we are loading the content...
Similar Documents
Realization of Boolean Functions Using Binary Pi-sigma Networks (1991)
| Content Provider | CiteSeerX |
|---|---|
| Author | Shin, Yoan Ghosh, Joydeep |
| Description | This paper introduces a higher-order neural network called the Binary Pi-sigma Network (BPSN), which is a feedforward network with a single "hidden" layer and product units in the output layer. As training proceeds, the BPSN forms an internal representation of the conjunctive normal form expression corresponding to the Boolean function to be learned. This enables the network to have a regular structure and to exhibit fast learning. We formally prove that the BPSN can realize any Boolean function. Simulation results show that the network converges very fast and in a stable manner. Introduction Since the introduction of the McCulloch-Pitts neuron, there have been many efforts to model logical expressions using neural networks [1]. The McCullochPitts neuron can be used as a threshold logic unit and hence can implement an AND or OR function of its inputs. Negation of the inputs also allows the NOT function. Thus, any Boolean expression can be realized using either the disjunctive normal ... Proceedings of Conference on Artificial Neural Networks in Engineering |
| File Format | |
| Language | English |
| Publisher | ASME Press |
| Publisher Date | 1991-01-01 |
| Access Restriction | Open |
| Subject Keyword | Fast Learning Binary Pi-sigma Network Threshold Logic Unit Feedforward Network Many Effort Single Hidden Layer Training Proceeds Higher-order Neural Network Conjunctive Normal Form Expression Mccullochpitts Neuron Boolean Expression Boolean Function Regular Structure Logical Expression Product Unit Stable Manner Output Layer Neural Network Simulation Result Mcculloch-pitts Neuron Internal Representation |
| Content Type | Text |
| Resource Type | Article |
