Learning Interpretations Using Sequence Classification
Proceedings of the Eleventh International Conference on Grammatical Inference, PMLR 21:220-223, 2012.
In this paper we present a system that assigns interpretations, in the form of shallow semantic frame descriptions, to natural language sentences. The system searches for relevant patterns, consisting of words from the sentences, to identify the correct semantic frame and associated slot values. For each of these choices, a separate classifier is trained. Each classifier learns the boundaries between different languages, which each correspond to a particular class. The different classifiers each have their own viewpoint on the data depending on which aspect needs to be identified.