MOQA - Meaning-Oriented Question-Answering

The paper describes a system which uses a fact repository (FR) consisting of instances of ontological concepts to answer sequences of questions. The fact repository is populated automatically by analyzing texts to produce complex text meaning structures. These are then mapped to fact repository instances and stored in a relational database. The documents used by the system are news articles in English, Arabic, and Farsi. The FR, the ontology and the lexicons for the three languages, together with the working memory that includes the intermediate results of the system operation, are the major static knowledge sources in the system. A multimedia interface allows queries to the system to be posed in a variety of ways including natural language. The user interacts with the system through input text, tables, generated text, maps, and relationship diagrams. The results of a query session are stored and can be reviewed and used to generate reports or to restart the session at any point in its history.