Natural language understanding system

A method for retrieving text comprising the steps of:

• inputting textual information to a natural language understanding module;
• translating, using said natural language understanding module, said textual information into a logic form which expresses objects and events in said textual information as concepts and relationships between said concepts that identify the structure expressed in said textual information;
• inputting a query to said natural language understanding module;
• translating, using said natural language understanding module, said query into a logic form which expresses objects and events in said query as concepts and relationships between said concepts that identify the structure expressed in said query;
• comparing said logic form of said query to said logic form of said textual information;
• and
• retrieving said textual information when said query's logic form is similar to said textual information's logic form.

A method for creating a cognitive model comprising the steps of:

• storing textual information in storage of said computer system;
• inputting said textual information to a natural language understanding (NLU) module, said NLU module comprising parsing, disambiguation, formal semantics, anaphora resolution and coherence modules that consult a naive semantic lexicon having at least one entry that identifies at least one sense of a word, said entry including a reference to an ontological classification network, syntactic information and a plurality of semantic properties;
• generating, using said natural language understanding module, a cognitive model which expresses objects and events in said textual information as concepts and relationships between said concepts that identify the structure expressed in said textual information;
• and
• generating a concept index of said concepts in said cognitive model, said concept index is used to locate the textual information associated with concepts in said cognitive model.

A method for retrieving text in a computer system comprising the steps of:

• storing in said computer system textual information;
• generating, using a natural language understanding module comprising parsing, disambiguation, formal semantics, anaphora resolution and coherence modules consulting a naive semantic lexicon having at least one entry that identifies at least one sense of a word, said entry including a reference to an ontological classification network, syntactic information and a plurality of semantic properties, a cognitive model of said textual information which expresses objects and events in said textual information as concepts and relationships between said concepts that identify the structure expressed in said textual information;
• generating, using said natural language understanding module, a cognitive model of a query which expresses objects and events in said query as concepts and relationships between said concepts that identify the structure expressed in said query;
• and
• generating a list of textual information relevant to said query based on said cognitive model of said textual information and said cognitive model of said query.

An article of manufacture comprising:

• a computer usable medium having computer readable program code embodied therein for interpreting natural language input comprising:
• computer readable program code configured to cause a computer to parse said natural language input driven by the content of said natural language input and expectations about said natural language input using a naive semantic lexicon having at least one entry that identifies at least one sense of a word, said entry including a reference to an ontological classification network, syntactic information, and a plurality of semantic properties;
• computer readable program code configured to cause a computer to disambiguate during said parse said natural language input using said naive semantic lexicon;
• computer readable program code configured to cause a computer to generate a discourse representation said natural language input using said naive semantic lexicon;
• computer readable program code configured to cause a computer to resolve references in said natural language input using said naive semantic lexicon;
• and
• computer readable program code configured to cause a computer to determine relationships that exist in said natural language input using said naive semantic lexicon.