Class Brain Coprocessor based on Neuromorphic Circuit for Efficient Non-Formalization and Unstructured Information Processing

Abstract

Class brain coprocessor is a type of coprocessor based on neuromorphic circuits that includes a memory module for storing training characteristics information, a processing module based on a hierarchical structure, and an encoder and decoder for input and output. This research proposes a memory module with a training characteristics storehouse and/or configurable training characteristics storehouse, and a processing module with a solidification functional network module and/or configurable functionality mixed-media network modules, which enhances the extended capability of the coprocessor. The proposed coprocessor employs distributed storage and concurrent collaborative processing, making it particularly suitable for handling non-formalization problems and unstructured information, as well as form problems and structured messages. The results show that this coprocessor significantly accelerates the speed of computers in processing class brain,informationificial intelligence, and reduces energy consumption while improving fault-tolerant ability, reducing programming complexity, and improving computing power.