AI-enabled systems

Decomposition method of functional characteristics of artificial intelligence systems

S. V. Garbuk

All-Russian Institute of Scientific and Technical Information RAS, Moscow, Russia Technical Committee for Standardization No. 164 ``Artificial Intelligence'' Federal Agency on Technical Regulating and Metrology, Moscow, Russia

Abstract: In making decisions on the possibility of using artificial intelligence systems to solve specific applications, one of the critical problems is an objective assessment of the functional characteristics for these systems in the planned operating conditions. The article proposes a universal approach to justifying the composition of environmental factors, the variability of which should be taken into account to ensure the required level of representativeness of the conducted tests of artificial intelligence systems. The proposed approach is based on the assumption that it is necessary to develop a classifier of elementary intellectual functions similar to the functional structure of human intelligence and having certain, unified within one class, requirements to the variability of their testing conditions.

Keywords: artificial intelligence, conformity assessment, intellectual functions, natural intelligence.

DOI: 10.14357/20718594250102

Bibliographic databases:

•