Every product from NeuroSystems includes KORTeX, its Artificial Intelligence engine based on multilayered artificial neural networks (ANN) able to use -when necessary- genetic algorithms for its own optimization. Usually, ANN are used to model complex relationships between inputs and outputs or to find patterns in data.

 

This systems has two learning paradigms, each corresponding to a particular abstract learning task. These are:

Supervised learning, to let the network converge to a precise final state.

Unsupervised learning, to let the network freely converge to any final state.

 

Also, KORTeX features many activation functions and a large number of types of neural networks, such as:

 

(for supervised learning)

ALN : Adaptive Logic Network

CasCor : Cascade Correlation

RCC : Recurrent Cascade Correlation

EKF : Extended Kalman Filter

REKF : Recurrent Extended Kalman Filter

PNN : Probabilistic Neural Network

GRNN : General Regression Neural Network

FCM : Fuzzy Cognitive Map

 

(for unsupervised learning)

ART1 : Binary Adaptive Resonance Theory

ART2 : Analog Adaptive Resonance Theory

DH : Discrete Hopfield

CH : Continuous Hopfield

 

Note that NeuroSystems' Augmented Reality technology also uses specific KORTeX functionnalities for Computer Vision.

 

To know more about Artificial Intelligence and neural networks, click here.