Matching syntactic-semantic graphs for semantic relation assignment

Abstract

We present a graph-matching algorithm for semantic relation assignment. The algorithm is part of an interactive text analysis system. The system automatically extracts pairs of syntactic units from a text and assigns a semantic relation to each pair. This is an incremental learning algorithm, in which previously processed pairs and user feedback guide the process. After each assignment, the system adds to its database a syntactic-semantic graph centered on the main element of each pair of units. A graph consists of the main unit and all syntactic units with which it is syntactically connected. An edge contains information both about syntax and about semantic relations for use in further processing. Syntactic-semantic graph matching is used to produce a list of candidate assignments for 63.75% of the pairs analysed, and in 57% of situations the correct relations is one of the system's suggestions; in 19.6% of situations it suggests only the correct relation.