Method for recognizing and interpreting patterns in noisy data sequences

• US 20100198592A1
• Filed: 01/27/2010
• Published: 08/05/2010
• Est. Priority Date: 02/02/2009
• Status: Active Grant

• Current Assignee: Jerry Lee Potter
• Original Assignee: Jerry Lee Potter
• Inventors: Potter, Jerry Lee
• Granted Patent: US 8,504,374 B2
• Time in Patent Office: Days
• US Class Current: 704/233
• 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 method for correctly interpreting a sequence of possibly noisy input data using a digital system, comprising:

(a) A digital system consisting of a plurality of digital devices and devices containing digital components such as but not limited to a computers, personal computers, laptops, personal digital assistants, cell phones, remote controls, inventory control devices, home appliances, automobile, robots, factory machines, construction machines, farming machines, airplanes, or remotely piloted aircrafts. Such devices may be interconnected by wired or wireless means to form the digital system and may operate in sequential or parallel mode; and

(b) Providing a receiving means and one or more digital devices of the digital system to receive;

(i) A plurality of top level input sequences of possibly noisy data. Said top level input sequences of data comprise a plurality of data types both noisy and not. Said top level sequence comprises a plurality of character, integer, real, single precision, double precision or quadruple precision data. Said top level input sequences may be organized into scalars, strings, arrays or any combination thereof; and

(ii) A plurality of homonym files and a means to modify and augment said homonym files. Said homonym files contain sufficient information to allow the correct alternative of the items in the homonym files to be substituted into the sequence of possibly noisy input data; and

(iii) A plurality of thematic pattern files and a means for augmenting and modifying said thematic pattern files; and

(iv) A plurality of data files constituting one or more databases; and

(c) A means for generating lexicons representing the semantic content of said data files; and

(d) Generating a plurality of alternative sequences of the top level possibly noisy input sequences using VP homonym file substitutions and assigning a score to the alternative sequences based on the nature of the substitutions; and

(e) Generating and scoring a plurality of said alternative sequence/thematic pattern pairs; and

(f) Generating and scoring a plurality of alternative segment parsings of the alternative sequences based on the alternative thematic patterns; and

(g) Generating and selecting the alternative sequence/alternative thematic pattern/alternative parsing triple frame that best maps the original top level possibly noisy input sequence onto one or more records of one or more of the said databases. Such records and the corresponding thematic pattern represent the interpretation of the said top level input sequence; and

(h) Verifying one or more sequences of symbols corresponding to the said interpretation for the correctness of the alternative sequence/alternative thematic pattern/alternative parsing frame mapping to the said databases; and

(i) Using said interpretation and feedback from the said verification operation to identify one or more actions to be issued. Said actions may be internal to the said digital system and used to modify, process, improve, educate, teach, update, search, interrogate, save and create databases. Said operations may be external to the said digital system and used to control other computers and/or digital devices. Said external operations may be in a plurality of modes of output;

oral, visual, odoriferous, electromechanical, verbal, command, feedback, textual, etc.; and

2. A method, as claimed in claim 1, wherein said means for generating lexicons representing the semantic content of said data files further comprises generating lexicon entries consisting of a system wide table, row and column identification such as a number, an index, etc. and the associated combination of table, row and column names. Said combinations include, but are not limited to:

(a) Row names,

(b) Column names,

(c) Row/Column names,

(d) Column/Row names

(e) Table/row names,

(f) Table/column names,

(g) Table/row/column names,

(h) Table/column/row names, and

(i) etc.; and

3. A method, as claimed in claim 1, wherein said generating and scoring a plurality of said alternative sequences/thematic pattern pairs further comprises:

(a) Determining the number of VP homonym changes in the alternative sequences; and

;

(b) Counting the number of the thematic patterns'"'"' terminal symbols that are matched by terminal symbols in the alternative sequences; and

(c) Determining the number of terminal symbols in the alternative sequences; and

(d) Determining the number of terminal symbols in the thematic patterns; and

(e) Determining the number of VP segments in the alternative sequences; and

(f) Determining the number of VP segments in the alternative thematic patterns; and

4. A method, as claimed in claim 1, wherein generating and scoring a plurality of alternative segment parsings of the alternative sequences based on the alternative thematic patterns step further comprises:

(a) Determining the positions of the alternative thematic patterns'"'"' terminal symbols in the alternative sequences so as to divide the said alternative sequences into verb segments, preposition segments and TRCC segments compatible with the syntactic parsings dictated by the terminal symbols of the thematic patterns; and

(b) Determining a plurality of measurements and scores to evaluate the alternative VP segments'"'"' fit to the alternative patterns'"'"' parsings; and

(c) Generating a plurality of alternative TRCC segments using sets of noun and adjective homonyms, near homonyms, multiple word sequences and similar substitutions from the TRCC homonym files; and

(d) Determining a plurality of measurements and scores to evaluate and select the best alternative TRCC segment mappings to one or more database lexicon entries; and

5. A method, as claimed in claim 4, wherein said determining a plurality of measurements and scores to evaluate the alternative VP segments'"'"' fit to alternative patterns'"'"' parsings step further comprises:

(a) Assigning basic and advanced scores to the VP segments of the alternative sequence/alternative thematic pattern/alternative parsing triple frame based in part on the degree of the match between the alternative sequences'"'"' VP segments and alternative parsings. The said scores'"'"' components are obtained in part, but not limited to;

(i) measuring the match between (1) the number of terminal symbols matched between the alternative sequence'"'"'s VP segments and the equivalent thematic pattern segments and (2) the length of the longest of the alternative sequence VP segments and the equivalent thematic pattern segments; and

(ii) measuring the match between the number of terminal symbols in the alternative sequence'"'"'s VP segments and the number of terminal symbols in the corresponding thematic pattern segments; and

(iii) measuring the match between (1) the number of terminal symbols matched between the alternative sequence'"'"'s VP segments and the equivalent thematic pattern segments and (2) the number of such matches that are in the correct (same) order; and

(iv) measuring the match between (1) the number of terminal symbols matched between the alternative sequence'"'"'s VP segments and the equivalent thematic pattern segments and (2) the number of such matched terminal symbols that are in the correct (same) position; and

6. A method, as claimed in claim 4, wherein determining a plurality of measurements and scores to evaluate and select the best alternative TRCC segment mappings to one or more database lexicon entries further comprises:

(a) assigning basic and advanced noun phrase scores to the TRCC segments of the alternative sequence/alternative thematic pattern/alternative parsing triple based on the degree of the match between the alternative TRCC segments and the database lexicons. The said scores'"'"' components are obtained in part but not limited to;

(i) measuring the match between the number of terminal symbols in common with the alternative TRCC segments and the lexicon segments scaled by the length of the longest of the said alternative noun phrase segment or the lexicon entry;

(ii) measuring the match between the number of terminal symbols in the alternative noun phrase segment and the lexicon entry;

(iii) measuring the number of VP terminal symbols that occur in the alternative TRCC segment; and

(b) using the TRCC scores to select the best match between alternative noun phrases and lexicon entries and thereby selecting a lexicon entry, a data file and a database; and

7. A method, as claimed in claim 1, wherein generating and selecting the alternative sequence/alternative thematic pattern/alternative parsing triple frame that best maps the original top level possibly noisy input sequence onto one or more records of one or more of the said databases further comprises:

(a) generating an alternative sequence frame score based on the verb, preposition and noun phrase scores to assess the quality of the match to the databases. The said alternative sequence/alternative thematic pattern/alternative parsing frame scores are obtained in part but not limited to;

(i) determining a basic frame score consisting of;

(1) the sum of the terminal symbols in the verb, preposition and noun phrase segments that were not matched; and

(2) the sum of the number of homonym changes in the verb, preposition and noun phrase segments; and

(ii) determining an advanced frame score consisting of;

(1) the sum of the individual advanced verb segment and preposition segment scores of all of the verb and preposition segments in the alternative sequence;

(2) the sum of the individual advanced noun phrase scores of all of the noun phrase segments in the alternative sequence;

(3) calculating the number of verb/preposition and noun phrase segments that had advanced scores greater than zero;

(4) calculating the number of noun phrases in the alternative sequences that were mapped onto a data file lexicon other than the current default data file lexicon; and

(iii) determining an adjusted advanced frame score by adding a factor to the advanced frame score based on the number of verb and prepositional homonym changes to the alternative input sequence; and

(b) selecting the alternative sequence/alternative thematic pattern/alternative parsing triple frame that best represents the interpretation of the possibly noisy input sequence by;

(i) selecting the alternative sequence/alternative thematic pattern/alternative parsing triple frame whose basic frame score is better than all other basic frame scores by a specified parameter;

or failing that

(ii) selecting the alternative sequence/alternative thematic pattern/alternative parsing triple frame whose basic frame score is better than or equal to all other basic frame scores by a specified parameter and its advanced frame score is better than all other advanced frame scores;

or failing that

(iii) selecting the alternative sequence/alternative thematic pattern/alternative parsing triple frame whose basic frame score is better than or equal to all other basic frame scores by specified parameter and its advanced frame score is better than or equal to all other advanced frame scores by a specified parameter and the adjusted advanced frame score is better than all other adjusted advanced frame scores;

or failing that

(iv) selecting the alternative sequence/alternative thematic pattern/alternative parsing triple frame with the best combination score.