Insight Grammar Learning

We report on computational experiments in which a learning agent incrementally acquires grammar from a tutoring agent through situated interactions. The learner is able to (i) detect impasses in routine language processing, such as missing a grammatical construction to integrate a word in the rest of the sentence structure, (ii) move to a meta-level to repair these impasses, primarily based on semantics, and (iii) then expand or restructure his grammar using insights gained from repairs. The paper proposes a cognitive architecture able to support this kind of insight learning and tests it with a computational implementation for a grammar learning task.