Invention:
This technology is a new model of computation leveraging the power of the networked, parallelized structure of neural networks and the basic operations of computers that are termed the "Queron". The Queron would be considered a generalization of the neuron and Queral Networks a generalization of computation and ANNs. Querons are capable of connecting and listening to other Querons based on their input requirements and will allow incoming connections from other Querons in the same manner. This will allow for massive parallelization currently absent from current solutions.
Background:
Traditionally Artificial Neural Networks (ANNs) are composed of interconnecting artificial neurons programmed to mimic the properties of biological neurons. The individual neurons in an ANN have been extensively investigated and ANNs have been developed which can perform very complex computations. Even so, current ANN models are not highly parallelized in the same way organic neural networks are and therefore are unable to match their computational power.
Advantages:
- Can perform very complex computations
- Highly parallelized in the same way organic neural networks are and therefore are able to match their computational power.
- Allow for massive parallelization currently absent from current solutions.
Applications:
- To create improved network models
- Understand the meaning of a text, so search engine optimizations on parallelized systems might be of interest
- The higher computing power of parallel networks can have implications for Computer Security and Cryptography
