Plausible Neural Networks (PNN) Technologies

Plausible Neural Networks is designed for the next generation of intelligent computing machines. The computation of PNN can be implemented by neuromorphic circuits and other kinds of physical computing systems. PNN has several innovative breakthrough technologies, which are under US and PCT patents.

 At the current stage of development, PNN is an intelligent system for data analysis with the following features.

• Self-organization - PNN learning is entirely unsupervised, based on the competitive neuronal signals for entropy reduction.

• Fast learning algorithms - PNN has the fastest learning algorithm among the current neural network and machine learning methods. 

• Universal data analysis - PNN performs clustering, classification, function estimation, associative memory and statistical inference.

• Handles different types of data - PNN employs unified coding scheme for cross-scale data analysis.

• Handles missing data - PNN performs analysis without the need of inputting missing data values.

• Confidence measures - PNN provides confidence measures for predictions.

• Features selection - PNN determines the most important attributes based on mutual information.

• Inference rules - PNN extracts rules for the relationship between variables through network queries.

 

 

Home- About -Product - News - Research - Technologies - Contact