We describe our research on automatically generating rich semantic annotations of text and making it available on the Semantic Web. In particular, we discuss the challenges involved in adapting the OntoSem natural language processing system for this purpose. OntoSem, an implementation of the theory of ontological semantics under continuous development for over fifteen years, uses a specially constructed NLP-oriented ontology and an ontological-semantic lexicon to translate English text into a custom ontology-motivated knowledge representation language, the language of text meaning representations (TMRs). OntoSem concentrates on a variety of ambiguity resolution tasks as well as processing unexpected input and reference. To adapt OntoSem’s representation to the Semantic Web, we developed a translation system, OntoSem2OWL, between the TMR language into the Semantic Web language OWL. We next used OntoSem and OntoSem2OWL to support SemNews, an experimental web service that monitors RSS news sources, processes the summaries of the news stories and publishes a structured representation of the meaning of the text in the news story.

We argue in favor of using a graph-based representation for language meaning and propose a novel learning method to map natural language text to its graph-based meaning representation. We present a grammar formalism, which combines syntax and semantics, and has ontology constraints at the rule level. These constraints establish links between language expressions and the entities they refer to in the real world. We present a relational learning algorithm that learns these grammars from a small representative set of annotated examples, and show how this grammar induction framework and the ontology-based semantic representation allow us to directly map text to graph-based meaning representations.

In this paper, we experiment with the task of contextual synonym expansion, and compare the benefits of combining multiple lexical resources using both unsupervised and supervised approaches. Overall, the results obtained through the combination of several resources exceed the current state-of-the-art when selecting the best synonym for a given target word, and place second when selecting the top ten synonyms, thus demonstrating the usefulness of the approach.