A system and method for facilitating machine interpretation of potentially ambiguous input, such as natural language. An example method includes constructing an ontology based on a database schema, wherein the ontology includes one or more indications of relationship strengths characterizing two or more entities of the ontology; collecting context information pertaining to a relevant domain; receiving the potentially ambiguous input; employing the input and the ontology to estimate a meaning of the input using the context information, the one or more indications of relationship strength, and an estimation of a probability that one or more portions of the natural language input refer to a particular entity.