This paper tries to emphasize some critical problems which appear in the main approaches of AI related to the development of advanced intelligent systems. The basis of the reasoning process is knowledge acquisition by learning, as an expression of the perception. The capability of perception is considered of being of the same importance as the reasoning for intelligent systems, especially for those that directly perform in the environment at whose events have to act in real time to reach the proposed goal. But also a reference system is needed the system activity to run in accordance with it. The aim of the proposed value_system-based architecture is that systems accomplish reflex-like actions as well as actions triggered by decision processes.
perception, reasoning, assessment, value_system, real-time