A semantic gap is the mismatch between data a computational system can process and the deeper real-world meaning that data represents. AI may recognize a handwritten 9 without understanding that it probably denotes a child’s age on a birthday card. section 7.0
The gap limits systems that learn mainly from statistical patterns. Recognition, language understanding, and game playing can succeed through pattern recognition and optimization, while constrained puzzles such as Farmer-Wolf-Goat-Cabbage require explicit rules, deduction, and forward planning. section 7.0
A number without its meaning
An image model can classify a handwritten symbol as 9, yet fail to infer whether it represents an age, a quantity, or something else. Closing this gap may require symbolic reasoning or smart agents that work beyond data-only learning. section 7.0
