Error correction using utterance disambiguation techniques

This paper describes a mechanism for identifying errors made by a student during a computer-aided language learning dialogue. The mechanism generates a set of ‘perturbations’ of the student’s original typed utterance, each of which embodies a hypothesis about an error made by the student. Perturbations are then passed through the system’s ordinary utterance interpretation pipeline, along with the student’s original utterance. An utterance disambiguation algorithm selects the best interpretation, performing error correction as a side-effect.