Representations and algorithms for cognitive learning

摘要

This is a report summarizing our progress towards a theory of cognitive learning. It is concerned with an algorithm that recognizes, selects and formulates in an internal language problems that arise in an external environment. This algorithm revises its representation of the environment and uses it to cope with self-selected problems.The algorithm depends on the formation of hypotheses and their use to select actions. The key ideas of this project are major new additions to a theory of representation of knowledge built on an inductive predicate logic.