Representing Syntactic-Semantic Knowledge from English Texts

We present a unified representational framework LSeN (Language to Semantic Network) for representing syntactic-semantic knowledge extracted from natural language (e.g. English) texts. Our representation is a variant of the standard semantic networks with some important distinctive features that facilitate bridging the interface between syntactic and semantic knowledge using automatic computational tools. We show the basic constructs in our representation and briefly describe the developed tool for automatically building the knowledge structure in LSeN from parsed English sentences. Our representational framework and the associated tool establish a good platform for further studies in automatic knowledge extraction from texts.