The Research On Uncertainty Of Artificial Intelligence

The uncertainty problem is an important aspect of artificial intelligence research,which objectively reflects the nature feature of the world. Traditional research onartificial intelligence has the issues of missing uncertainty problem, hindering furtherdevelopment of artificial intelligence.This thesis chooses the uncertainty problem in artificial intelligence as the studyobject. First, we elaborated that uncertainty reflects the continuity of time and spacefrom a philosophical point of view. Then, we carried out an in-depth study on theuncertainty problem in artificial intelligence from four aspects: logic basis, researchmethod, technology development, and application prospect. On the logical basis ofartificial intelligence, we argue that the development of uncertainty-inclusive flexiblelogic will play a catalytic role for artificial intelligence; on the research methods ofartificial intelligence, we supplement and perfect the widely accepted mechanismtheory and try to put forward that the introduction of uncertain state into the hardware,resulting in {0,1,α} tri-state model, and the consideration of social factors into softwaresimulation will be important directions for the development of new intelligent machines;on the technology development of artificial intelligence, we argue that the uncertainty ofinner execution mechanism leads to the difficulties of development and the uncertaintyof the external environment leads to difficulties of process control; on the applicationprospects of artificial intelligence, we introduce that the development of artificialintelligence may change the social relationship and encounter ethical problem. Thisprovides a new idea for the development of artificial intelligence.