Topological inference of teleology: Deriving function from structure via evidential reasoning

摘要

Reasoning about the physical world is a central human cognitive activity. One aspect of such reasoning is the inference of function from the structure of the artifacts one encounters. In this article we present the Topological iNference of Teleology (TnT) theory, an efficient means of inferring function from structure. TnT comprises a representation language for structure and function that enables the construction, extension, and maintenance of the domain-specific knowledge base required for such inferences, and an evidential reasoning algorithm. This reasoning algorithm trades deductive soundness for efficiency and flexibility. We discuss the representations and algorithm in depth and present an implementation of TnT, in a system called Carnot. Carnot demonstrates quadratic performance and broad coverage of the domain of single-substance thermodynamic cycles, including all such cycles presented in a standard text on the subject. We conclude with a discussion of Carnot-based coaching tools that we have implemented as part of our publicly available CyclePad system, which is a design-based learning environment for thermodynamics.