Automated comprehension of natural language via constraint-based processing

• US 10,599,775 B2
• Filed: 03/21/2018
• Issued: 03/24/2020
• Est. Priority Date: 04/18/2014
• Status: Active Grant

• Current Assignee: Thomas A. Visel
• Original Assignee: Thomas A. Visel
• Inventors: Visel, Thomas A.
• Primary Examiner(s): Yehl, Walter
• Application Number:
  US15/927,234
• Publication Number:
  US 20180217982A1
• Time in Patent Office: 734 Days
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A consistent meaning framework (CMF) graph including a plurality of nodes linked by a plurality of edges is maintained in data storage of a data processing system. Multiple nodes among the plurality of nodes are meaning nodes corresponding to different word meanings for a common word spelling of a natural language. Each of the multiple word meanings has a respective one of a plurality of associated constraints. A natural language communication is processed by reference to the CMF graph. The processing includes selecting, for a word in the natural language communication, a selected word meaning from among the multiple word meanings based on which of the plurality of associated constraints is satisfied by the natural language communication. An indication of the selected word meaning is stored in data storage.

Claims

1. A method in a data processing system of generating a response to a natural language communication, the method comprising:

prior to receipt by the data processing system of a natural language communication for processing, maintaining, in data storage of the data processing system, a consistent meaning framework (CMF) graph providing word meanings for various word spellings of a natural human language, wherein the CMF graph includes a plurality of nodes linked by a plurality of edges, and wherein the plurality of nodes includes;

a plurality of meaning nodes each specifying a respective word spelling, wherein each of the plurality of meaning nodes is linked by a respective one of the plurality of edges to another meaning node serving as a parent node indicating what the word spelling means, and wherein at least two of the plurality of meaning nodes correspond to different word meanings for a common word spelling of the natural human language; and

a plurality of sense nodes each linked to a respective one of the plurality of meaning nodes by one of the plurality of edges, each of the plurality of sense nodes having a respective associated sense defined by a natural language textual description indicating a word meaning to which the linked meaning node among the plurality of meaning nodes refers, such that different word meanings of the common word spelling can be disambiguated;

in response to receipt by the data processing system of the natural language communication, a processor of the data processing system processing the natural language communication by reference to the CMF graph, wherein the processing includes parsing the natural language communication and selecting, for each of multiple word spellings in the natural language communication, a selected word meaning from among the word meanings provided by the CMF graph; and

based on results of the processing, the processor outputting a natural language response to the natural language communication.

2. The method of claim 1, and further comprising:

maintaining, in data storage, a words map of word spellings, wherein the words map indexes into each of the plurality of meaning nodes in the CMF graph.

3. The method of claim 1, wherein:

multiple of the word meanings are each associated by the CMF graph with a respective one of a plurality of constraints defining a grammatical class of word meaning and indicating a sense of word meaning; and

the selecting includes selecting word meanings from among the word meanings provided by the CMF graph based on which of the plurality of associated constraints is satisfied by the natural language communication.

4. The method of claim 3, wherein the selecting includes determining whether or not one of the associated constraints is satisfied by a root word of a word spelling appearing in the natural language communication.

5. The method of claim 3, wherein the selecting includes determining whether or not one of the associated constraints is satisfied by at least one of a set including testing a word meaning for membership in a specified class, testing a word meaning for compatibility with a linguistic role, and testing a sense of a word meaning.

6. The method of claim 3, wherein the parsing includes parsing the natural language communication utilizing constraints to identify phrasal structures in the natural language communication and recording the phrasal structures in data storage.

7. The method of claim 3, wherein the processing includes parsing the natural language communication in a direction of the natural human language until no additional constraints are satisfied and then parsing the natural language communication in a reverse direction.

8. The method of claim 3, wherein the parsing includes applying multiple nested constraints to determine the selected word meaning for at least one word spelling in the natural language communication.

9. The method of claim 8, wherein the applying includes applying a first constraint for a first word spelling and applying a second constraint for a second word spelling, wherein both of the first and second constraints invoke a third constraint.

10. The method of claim 8, wherein the applying includes applying a first constraint to a multi-word phrase of the natural language communication and applying a second constraint to an individual word spelling in the multi-word phrase.

11. The method of claim 10, wherein:

the first constraint defines grammatical structure of the multi-word phrase; and

the second constraint qualifies a part of speech of the individual word spelling.

12. The method of claim 1, and further comprising:

prior to the processing, tokenizing the natural language communication, wherein the tokenizing includes;

allocating, in data storage, a respective token storage area for each of the multiple word spellings in the natural language communication; and

storing in a given token storage area a reference to a word meaning and a semantic data type determined via constraint-based parsing.

13. The method of claim 1, wherein outputting the natural language response includes presenting an audible natural language response via a text-to-speech converter.

14. The method of claim 1, and further comprising the data processing system performing semantic analysis on the natural language communication and changing which of the plurality of meaning nodes are linked by at least one of the plurality of edges in the CMF graph based on the semantic analysis of the natural language communication.

15. The method of claim 14, wherein performing semantic analysis includes evaluating one or more constraints utilizing one or more constraint handlers.

16. The method of claim 1, and further comprising:

based on the parsing, the processor performing semantic analysis of at least a sentence fragment of the natural language communication to obtain at least mid-level comprehension information regarding the sentence fragment; and

the processor determining the natural language response based at least in part on the mid-level comprehension information.

17. The method of claim 1, wherein the natural language communication is in an input natural human language and the natural language response is in a different target natural human language.

18. The method of claim 1, and further comprising the processor determining the natural language response utilizing an artificial intelligence (AI) system.

19. The method of claim 1, wherein the natural language response includes at least one of a set including an indication of emotion evoked by the natural language communication and an indication of emotion implied by a communicator of the natural language communication.

20. The method of claim 1, wherein the natural language response includes at least one of a set including a topic of the natural language communication, a summary of the meaning of the natural language communication, and a summary of classical linguistic structure of the natural language communication.

21. The method of claim 1, wherein the natural language response includes an assessment of an intent of a communicator of the natural language communication.

22. The method of claim 1, wherein the natural language response includes a definition of one or more word spellings in the natural language communication.

23. A program product for generating a response to a natural language communication, the program product comprising:

a storage device; and

program code stored within the storage device, wherein the program code, when executed by a processor of a data processing system, causes the data processing system to perform;

prior to receipt by the data processing system of a natural language communication for processing, maintaining, in data storage of the data processing system, a consistent meaning framework (CMF) graph providing word meanings for various word spellings of a natural human language, wherein the CMF graph includes a plurality of nodes linked by a plurality of edges, and wherein the plurality of nodes includes;

a plurality of meaning nodes each specifying a respective word spelling, wherein each of the plurality of meaning nodes is linked by a respective one of the plurality of edges to another meaning node serving as a parent node indicating what the word spelling means, and wherein at least two of the plurality of meaning nodes correspond to different word meanings for a common word spelling of the natural human language; and

a plurality of sense nodes each linked to a respective one of the plurality of meaning nodes by one of the plurality of edges, each of the plurality of sense nodes having a respective associated sense defined by a natural language textual description indicating a word meaning to which the linked meaning node among the plurality of meaning nodes refers, such that different word meanings of the common word spelling can be disambiguated;

in response to receipt by the data processing system of the natural language communication, a processor of the data processing system processing the natural language communication by reference to the CMF graph, wherein the processing includes parsing the natural language communication and selecting, for each of multiple word spellings in the natural language communication, a selected word meaning from among the word meanings provided by the CMF graph; and

based on results of the processing, the processor outputting a natural language response to the natural language communication.

24. A data processing system for generating a response to a natural language communication, the data processing system comprising:

a processor;

a storage device coupled to the processor; and

program code stored within the storage device, wherein the program code, when executed by the processor, causes the data processing system to perform;

prior to receipt by the data processing system of a natural language communication for processing, maintaining, in data storage of the data processing system, a consistent meaning framework (CMF) graph providing word meanings for various word spellings of a natural human language, wherein the CMF graph includes a plurality of nodes linked by a plurality of edges, and wherein the plurality of nodes includes;

a plurality of meaning nodes each specifying a respective word spelling, wherein each of the plurality of meaning nodes is linked by a respective one of the plurality of edges to another meaning node serving as a parent node indicating what the word spelling means, and wherein at least two of the plurality of meaning nodes correspond to different word meanings for a common word spelling of the natural human language; and

a plurality of sense nodes each linked to a respective one of the plurality of meaning nodes by one of the plurality of edges, each of the plurality of sense nodes having a respective associated sense defined by a natural language textual description indicating a word meaning to which the linked meaning node among the plurality of meaning nodes refers, such that different word meanings of the common word spelling can be disambiguated;

in response to receipt by the data processing system of the natural language communication, a processor of the data processing system processing the natural language communication by reference to the CMF graph, wherein the processing includes parsing the natural language communication and selecting, for each of multiple word spellings in the natural language communication, a selected word meaning from among the word meanings provided by the CMF graph; and

based on results of the processing, the processor outputting a natural language response to the natural language communication.