Abstract
Application of neuromorphic structures in various spheres of human activity on the basis of generalized neural elements will become possible if effective methods for verifying realizability of the logic algebra functions by one neuron element with a generalized threshold activation function and synthesis of such elements with a large number of edntries are developed. A notion of nucleus of Boolean functions in relation to a given system of characters was introduced and algebraic structure of nuclei and reduced nuclei of Boolean functions was investigated. Relation between the nuclei of the logic algebra functions which are realized by one generalized neural element and matrices of tolerance was established. It was shown that the Boolean function is realized by one generalized neuron element if and only if the nucleus of this function admits representation by the matrices of tolerance. If there is no nucleus relative to a specified system of characters for a Boolean function, then such a function is not realized by one generalized neural element in relation to a specified system of characters. On the basis of the properties of the matrices of tolerance, a number of necessary and sufficient conditions for realization of the logic algebra functions by one generalized neural element were obtained. Based on the sufficient conditions, an algorithm for synthesis of integer-valued generalized neural elements with a large number of entries was constructed. In the synthesis of integer-valued generic neural elements for realization of the logic algebra functions, a block representation of the Boolean function nucleus was used and based on the properties of the matrices of tolerance, coordinates of the integer vector of the structure of the generalized neural element were sequentially found
Highlights
Intensification of theoretical development and practical applications has been observed recently in the field of information technology and neurocomputers
We consider conditions sufficient for realizeability of Boolean functions by one generalized neural element which can be successfully used in synthesis of neural networks based on the generalized neural elements (GNE) with integer-valued structure vectors
Based on the results presented in the paper on the structure of nuclei and extended nuclei of Boolean functions with respect to the system of characters and properties of the matrices of tolerance, the following was established:
Summary
Intensification of theoretical development and practical applications has been observed recently in the field of information technology and neurocomputers This is due to the increased interest in information systems and neurolike structures that have found wide application in encription, protection of information, image recognition, forecasting and other fields of human activities. It should be mentioned that an extremely important requirement to the new methods of synthesis of generalized neural elements is that these methods should be practically suitable for synthesizing GNE with a large number of inputs. This is explained by the fact that the volume of information and the degree of complexity of the tasks that are solved in the neurobase are constantly growing. The application of generalized neural elements can reduce the number of artificial neurons in the neural networks employed for the tasks on recognition, compression and encoding of discrete signals and images
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: Eastern-European Journal of Enterprise Technologies
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
