Backward chaining

Backward chaining (or backward reasoning) is an inference method that can be described (in lay terms) as working backward from the goal(s). It is used in automated theorem provers, inference engines, proof assistants and other artificial intelligence applications.