
Quantitative symbolic process models are proposed as a high-level model representation for probabilistic causal knowledge and know-how. These representations have suitable expressive power and mathematical foundation to serve as an abstraction of intelligence in much of AI research.