Systems and methods for natural language communication with a computer

• US 8,271,264 B2
• Filed: 04/30/2009
• Issued: 09/18/2012
• Est. Priority Date: 04/30/2008
• Status: Active Grant

• Current Assignee: Deep Sky Concepts, Inc.
• Original Assignee: Glace Holding, LLC
• Inventors: Gorman, John G., Gorman, John W. R.
• Primary Examiner(s): Neway, Samuel G
• Application Number:
  US12/669,454
• Publication Number:
  US 20100211379A1
• Time in Patent Office: 1,237 Days
• Field of Search: 704/9
• US Class Current: 704/9
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A system for natural language communication with a computer, comprising a database configures to store at least one table the table configured to store associations between concepts in a text string and associated concept identifiers; and a builder interfaced with the database, the builder configured to construct a concept defined by a text string by performing a process comprising the following steps: assigning a unique identifier to the concept, breaking the concept into sub-concepts, assigning a unique identifier to each sub-concept, using legal joins to join sub-concepts into higher order sub-concepts, assigning a unique identifier to the higher-order concepts, and repeating the steps until the higher order sub-concepts join to form the concept, and storing the concept, sub-concepts, higher order sub-concepts and unique identifiers in a table in the database.

Claims

1. A system for knowledge representation in a computer that can enable natural language communication with a computer, the system comprising:

at least one hardware processor;

at least one database configured to store at least one table, the at least one table configured to store associations between concepts in a body of text and associated concept identifiers; and

a builder interfaced with the database, the builder configured to, when executed by the at least one hardware processor, construct a table that represents knowledge defined by a body of text comprising words and punctuation, the construction of the table comprising

assigning a unique identifier to an observation concept defined by the body of text,

building a tree comprising a plurality of sub-concepts by

breaking the observation concept into sub-concepts by identifying a subject concept, a seam concept, and a relative concept, and

joining sub-concepts into higher order sub-concepts until two or more sub-concepts join to form the concept, wherein said joining comprises

constructing a binary concept by either joining the subject concept and the seam concept or joining the seam concept and the relative concept, and

constructing a first higher order sub-concept by joining either the relative concept or the subject concept with the binary concept,

assigning unique identifiers to each of the plurality of sub-concepts including the subject concept, the seam concept, the relative concept, the binary concept, and the first higher order sub-concept, wherein the unique identifiers assigned to the observation concept and each of the plurality of sub-concepts are generated according to the same scheme, and

storing each of the concept and plurality of sub-concepts in a row in the at least one table, wherein each row comprises the concept or sub-concept, the unique identifier assigned to the concept or sub-concept, and the unique identifiers assigned to two or more lower order sub-concepts, if any, which were joined to form the concept or sub-concept;

wherein the plurality of sub-concepts comprise at least each of the words and punctuation in the body of text.

2. The system of claim 1, wherein the construction of the table further comprises constructing a pre-modifier concept and assigning a unique identifier to the pre-modifier concept, and forming a second higher order sub-concept by joining the pre-modifier concept with the first higher order sub-concept and assigning a unique identifier to the second higher order sub-concept.

3. The system of claim 1, wherein the construction of the table further comprises constructing a post-modifier concept and assigning a unique identifier to the post-modifier concept, and forming a second higher order sub-concept by joining the post-modifier concept with the first higher order sub-concept and assigning a unique identifier to the second higher order sub-concept.

4. The system of claim 1, wherein the unique identifiers assigned to the observation concept and plurality of sub-concepts comprise one or more of numeric identifiers, alphanumeric identifiers, row numbers, random numbers, acronyms, words, or letters.

5. The system of claim 1, wherein the builder is configured to construct a unique identifier from a first letter of each word comprising a concept or sub-concept.

6. The system of claim 1, wherein the builder is configured to construct a unique identifier from a word comprising a sub-concept.

7. The system of claim 1, wherein the builder is further configured to form higher order sub-concepts by joining the sub-concepts that represent the words and punctuation.

8. The system of claim 1, wherein each row of the at least one table further comprises metadata associated with a corresponding concept or sub-concept.

9. The system of claim 8, wherein the metadata comprises a vocabulary representation, if applicable, and a textual representation of the associated concept or sub-concept.

10. The system of claim 1, wherein each row of the at least one table further comprises a field comprising a response, if any, associated with the corresponding concept or sub-concept.

11. The system of claim 1, wherein the builder is configured to form higher order sub-concepts by receiving a drag-and-drop action, which comprises dragging a concept from one row and dropping it onto a concept in another row.

12. The system of claim 1, wherein the body of text-comprises one or more of a paragraph, chapter, book, article, or publication, and wherein the at least one table comprises a plurality of rows corresponding to observation concepts constructed by the builder to represent the knowledge represented in the body of text.

13. The system of claim 1, wherein the body of text describes an item or product, and wherein the at least one table constructed by the builder to represent the item or product can be used to tag the item.

14. A system for knowledge representation in a computer that can enable natural language communication with a computer, the system comprising:

at least one hardware processor;

a user interface configured to receive an input text string;

at least one database including at least one table representing at least one concept tree,

wherein the concept tree comprises a plurality of concepts, the plurality of concepts comprising a single root concept, a plurality of elementary concepts comprising words and punctuation, and joins between two or more of the plurality of elementary concepts,

wherein the plurality of concepts further comprise a subject concept, a seam concept, a relative concept, a binary concept, and a higher-order concept,

wherein the binary concept comprises a loin between either the subiect concept and the seam concept or the seam concept and the relative concept, and

wherein the higher-order concept comprises a join between the binary concept and either the relative concept or the subiect concept,

wherein the at least one table comprises a row for each of the plurality of concepts, each row comprising the concept, a unique identifier associated with the concept, and unique identifiers associated with two or more lower order concepts, if any, which form the concept, and

wherein the unique identifiers associated with each of the plurality of concepts are according to the same scheme; and

a parser interfaced with the at least one database and the user interface, the parser configured to, when executed by the at least one hardware processor, identify a concept represented by the text string, the identification of the concept comprising

breaking the text string into separate elements of language, the elements including words and punctuation,

using the separate elements of language to form an ordered string of concepts, and

forming higher order concepts comprising two lower order concepts, by searching the at least one table for joins between concepts, until the concept represented by the text string is formed.

15. The system of claim 14, wherein each row of the at least one table further comprises a field comprising a response, if any, associated with the corresponding concept, and wherein the system further comprises a responder configured to carry out instructions specified by the response.

16. The system of claim 15, wherein the responder is configured to output a text string to a display via the user interface as instructed by the response.

17. The system of claim 15, wherein the responder is configured to output an audio stream via the user interface as instructed by the response.

18. The system of claim 15, wherein the responder is configured to open a file or run a computer program as instructed by the response.

19. The system of claim 15, wherein the responder is configured to access a web site as instructed by the response.

20. The system of claim 14, wherein the user interface comprises a speech-to-text converter, and wherein the user interface is configured to generate the input text string from a speech input.

21. The system of claim 17, wherein the user interface comprises a text-to-speech converter and wherein the user interface is configured to generate the audio stream from a text string received from the responder.