Research on the Embedded Mathematical Model of Artificial Intelligence Measurement

In view of the unclear concept and weak basic theory of artificial intelligence, the bottleneck of basic measurement technology has not been solved. The relationship between AI metrics and basic cognitive quantity is extremely complex, so it is challenging and uncertain to scientifically construct its functional relationship. This article starts from revealing the internal mechanism, measurement principle and essential characteristics of artificial intelligence, the embedded physical laws are characterized, and the measurement model theory based on intelligence entropy is proposed. Multi-channel exploration of artificial intelligence measurement traceability methods provides ideas for forming the relationship model between artificial intelligence measurement indicators and benchmark quantities. The design verification method and approach of artificial intelligence measurement are established to ensure the reproducibility, accuracy and consistency of artificial intelligence measurement method and data.