An Optimal Assessment of Natural Language Student Input Using Word-to-Word Similarity Metrics

Abstract

We address in this paper the important task of assessing natural language student input in dialogue-based intelligent tutoring systems. Student input, in the form of dialogue turns called contributions must be understood in order to build an accurate student model which in turn is important for providing adequate feedback and scaffolding. We present a novel, optimal semantic similarity approach based on word-to-word similarity metrics and compare it with a greedy method as well as with a baseline method on one data set from the intelligent tutoring system, AutoTutor.