Reflection of Measurable Features in an Artificial Intelligence System

Key Points

• This research aims to improve the reliability of hypotheses generated by artificial intelligence systems by addressing uncertainty in measured features.
• Proposed a framework for using probability distributions to represent measured characteristics.
• Explored different formalization methods for numerical features within AI systems.
• Considered extended representations for input features in frames.
• Demonstrated that applying probability distributions increases reliability of hypotheses in AI systems.
• Highlighted various formalization techniques that can be utilized for input sequences of numerical features.