A New XML-Based Language for Neural Solution Interchange

This article introduces a framework for interchange of trained neural network models. An XML-based language (Neural Network Markup Language) for the neural network model description is offered. It allows to write down all components of neural network model, which are necessary for its reproduction. We propose to use XML notation for full description of neural models, including data dictionary, properties of training sample, preprocessing methods, details of network structure and parameters, method for network output interpretation.