Weighted Semantic Parsing: A Robust Approach To Interpretation Of Natural Language Queries

This paper focuses on a grammar-based approach to semantic interpretation, which combines the notions of robust and weighted parsing. In restricted domains of application in information extraction and natural language speech based information systems this approach shows acceptable performance. We present an overview of our research carried out within the recent project ISIS (Interaction through Speech with Information Systems) where techniques based on the above notions have been applied.