Method for recognizing and interpreting patterns in noisy data sequences

• US 8,504,374 B2
• Filed: 01/27/2010
• Issued: 08/06/2013
• Est. Priority Date: 02/02/2009
• Status: Active Grant

• Current Assignee: Jerry Lee Potter
• Original Assignee: Jerry Lee Potter
• Inventors: Potter, Jerry Lee
• Primary Examiner(s): Opsasnick, Michael N
• Application Number:
  US12/657,770
• Publication Number:
  US 20100198592A1
• Time in Patent Office: 1,287 Days
• Field of Search: 704/4, 704/254, 704/277
• US Class Current: 704/277
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence

Abstract

This invention maps possibly noisy digital input from any of a number of different hardware or software sources such as keyboards, automatic speech recognition systems, cell phones, smart phones or the web onto an interpretation consisting of an action and one or more physical objects, such as robots, machinery, vehicles, etc. or digital objects such as data files, tables and databases. Tables and lists of (i) homonyms and misrecognitions, (ii) thematic relation patterns, and (iii) lexicons are used to generate alternative forms of the input which are scored to determine the best interpretation of the noisy input. The actions may be executed internally or output to any device which contains a digital component such as, but not limited to, a computer, a robot, a cell phone, a smart phone or the web. This invention may be implemented on sequential and parallel compute engines and systems.

Claims

1. A system that receives a noisy input sequence, and converts that noisy input sequence into an action for a computing device comprising:

a processor that;

receives the noisy input sequence,

generates all possible combinations of alternative input sequences through homonymization of verbs and preposition (VP) homonyms which are stored,

pairs the alternative input sequences with all matching thematic patterns,

scores the VP homonym and thematic pattern combinations for the noisy input sequence by counting a number of keywords in common with each VP homonym and thematic pattern based on a stored table having stored therein thematic relation patterns for an environmental model,

generates a frame score for all VP homonym, thematic pattern and alternative frame parsing combinations which will be used for selecting a best match combination,

selects a best frame based on the frame score, and

interprets and initiates the action associated with the thematic pattern of the selected best frame on the computing device.

2. The system of claim 1, wherein the processor further:

generates, for each VP homonym and thematic pattern combination, a vector of frames of the possible VP homonym parsings based on the thematic pattern.

3. The system of claim 2, wherein a best match between pattern segments and input segments is evaluated and verified.

4. The system of claim 2, wherein the frames are used to parse the VP homonyms into VP and TRCC (Table, Row, Column, and Cell) segments, wherein the verb, prepositions and TRCC segments are processed producing a frame score that measures a fit of the frame'"'"'s VP homonym to the frame'"'"'s associated thematic patterns.

5. The system of claim 2, wherein the processor further determines a quality of a match between non-terminals and noun phrase segments.

6. The system of claim 4, wherein the VP homonym'"'"'s TRCC sequences are normalized and then homonynized with user specified character data, common mispronunciations, misrecognitions, true homonym substitutions for the table, row and column names and true homonym substitutions for constants.

7. The system of claim 4, wherein an adjusted advanced frame score includes a factor for a number of homonym changes to the VP homonyms.

8. The system of claim 1, wherein stored lexicons represent semantic content of environment data files and further include lexicon entries comprising one or more of a constant, string, word sequence, regular expression, image, sensor input, computer input or an environment wide table, row and column identification and an associated combination of table, row and column names, and wherein the associated combination of table, row and column names includes one or more of:

Table names, Row names, Column names, Row/Column names, Column/Row names, Table/row names, Table/column names, Table/row/column names, and Table/column/row names.

9. The system of claim 1, wherein the computing device is associated with one or more of a computer, an embedded computer, a personal computer, a laptop, a personal digital assistant, a cell phone, a remote control, an inventory control device, a home appliance, an automobile, a robot, a factory machine, a construction machine, a farming machine, an airplane, a sensor, and a remotely piloted aircraft.

10. A non-transitory computer readable storage medium, having stored thereon instructions, that when executed by the processor of claim 1, cause the steps therein to be performed.

11. A method to receive a noisy input sequence and convert that noisy input sequence into an action for a computing device comprising:

receiving the noisy input sequence,

generating, by a processor associated with the computing device, all possible combinations of alternative input sequences through homonymization of verbs and preposition (VP) homonyms which are stored,

pairing the alternative input sequences with all matching thematic patterns,

scoring the VP homonym and thematic pattern combinations for the noisy input sequence by counting a number of keywords in common with each VP homonym and thematic pattern based on a stored table having stored therein thematic relation patterns for an environmental model,

generating a frame score for all VP homonym, thematic pattern and alternative frame parsing combinations which will be used for selecting a best match combination,

selecting a best frame based on the frame score, and

interpreting and initiating the action associated with the thematic pattern of the selected best frame on the computing device.

12. The method of claim 11, further comprising

generating, for each VP homonym and thematic pattern combination, a vector of frames of the possible VP homonym parsings based on the thematic pattern.

13. The method of claim 12, wherein a best match between pattern segments and input segments is evaluated and verified.

14. The method of claim 12, wherein the frames are used to parse the VP homonyms into VP and TRCC (Table, Row, Column, and Cell) segments, wherein the verb, prepositions and TRCC segments are processed producing a frame score that measures a fit of the frame'"'"'s VP homonym to the frame'"'"'s associated thematic patterns.

15. The method of claim 12, further comprising determining a quality of a match between non-terminals and noun phrase segments.

16. The method of claim 14, wherein the VP homonym'"'"'s TRCC sequences are normalized and then homonynized with user specified character data, common mispronunciations, misrecognitions, true homonym substitutions for the table, row and column names and true homonym substitutions for constants.

17. The method of claim 14, wherein an adjusted advanced frame score includes a factor for a number of homonym changes to the VP homonyms.

18. The method of claim 11, wherein stored lexicons represent semantic content of environment data files and further include lexicon entries comprising one or more of a constant, string, word sequence, regular expression, image, sensor input, computer input or an environment wide table, row and column identification and an associated combination of table, row and column names, and wherein the associated combination of table, row and column names includes one or more of:

Table names, Row names, Column names, Row/Column names, Column/Row names, Table/row names, Table/column names, Table/row/column names, and Table/column/row names.

19. The method of claim 11, wherein the computing device is associated with one or more of a computer, an embedded computer, a personal computer, a laptop, a personal digital assistant, a cell phone, a remote control, an inventory control device, a home appliance, an automobile, a robot, a factory machine, a construction machine, a farming machine, an airplane, a sensor, and a remotely piloted aircraft.

20. The method of claim 11, wherein thematic relation patterns can be modified by a user and a user can add new thematic relation patterns to the environment.

21. The method of claim 13, wherein the verification allows a user to override the best match.