Learning by Understanding Analogies
This research describes a method for learning by analogy — i.e., for proposing new conjectures about a target analogue based on facts known about a source analogue. We formalize this process as a rule of plausible inference and present heuristics which guide this process towards efficiently postulating useful new conjectures. The most important rule involves the use of abstractions — abstract relations that encode solution methods to past problems.
Unable to display preview. Download preview PDF.