Vision virtual intelligent system

Abstract

The invention relates to a vision virtual intelligent system which comprises an input part, a processing part, a virtual intelligent core part, an output part and a knowledge base. The input part comprises a vision information processing module, a keyboard information processing module and an input combination module. The processing part comprises an input correlation module and a nonlinear solution module. The virtual intelligent core part comprises a central control module and a system basic module. The output part comprises an output conversion module. The knowledge base is divided into a data knowledge base and a structure knowledge base. The system is wide in purpose and wide in application range, and meanwhile has the advantages of automatically finding the solution scheme, intelligently optimizing itself and the like, the capacity of the system can be improved step by step along with the increase of use time, and the system can be applied to assisting in learning and living, can be used for the frontier field of unmanned driving and intelligent robots and the like, and is suitable for most occasions with images as main input and high requirements for system intelligence.

Description

technical field [0001] The invention relates to automatic learning and improvement of software, and relates to the fields of computer technology and artificial intelligence. Background technique [0002] With the rapid development of computers today, both hardware and software of computers have been fully developed. As the computing power of computers increases and our dependence on them increases, so do our demands on software. First of all, we require that the software can solve problems correctly or more accurately, and at the same time, we also hope that the software can solve more problems, or even all problems. Although nowadays, almost most of the needs can be met by finding different software, but the time cost and money cost brought by constantly looking for new software and learning new software are too high for people to bear. At the same time, our demands on software make programmers' jobs harder. They need to constantly modify the errors that appear in the so...

AI Technical Summary

Problems solved by technology

[0004] Existing artificial intelligence theories such as artificial neural networks can achieve self-learning to a certain extent, but they cannot self-adjust structural errors. At the same time, their learning efficiency and other issues are limited by the network structure itself.

Such a system has an upper limit, and the upper limit is often not high under existing theories

Images