Systems for natural language processing of sentence based queries

• US 20100005081A1
• Filed: 09/14/2009
• Published: 01/07/2010
• Est. Priority Date: 11/12/1999
• Status: Abandoned Application

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Bennett, Ian M.
• Application Number:
  US12/559,347
• Publication Number:
  US 20100005081A1
• Time in Patent Office: Days
• US Class Current: 707/4
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 40/216 using statistical methods
• CPC Class Codes: G06F 40/237 Lexical tools
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/42 Data-driven translation
• CPC Class Codes: G06F 40/44 Statistical methods, e.g. probability models
• CPC Class Codes: G09B 5/04 with audible presentation of the material to be studied sound recording or reproducing G11B
• CPC Class Codes: G09B 7/00 Electrically-operated teaching apparatus or devices working with questions and answers mechanically operated G09B3/00; computing arrangements G06F
• CPC Class Codes: G10L 15/142 Hidden Markov Models [HMMs]
• CPC Class Codes: G10L 15/18 using natural language modelling
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/285 Memory allocation or algorithm optimisation to reduce hardware requirements
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: H04M 2250/74 with voice recognition means voice activated dialling H04M1/271; voice recognition algorithms G10L15/00
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

Systems are disclosed for responding to text-based queries. The systems use natural language routines to identify word phrases present in a user query. The natural language routines may also consider the context of the user query, and may load one or both of appropriate grammars or dictionaries based on the context. A query formulation routine converts the words of the user query and word phrases generated by the natural language routines into a structured, combined query including search predicates for locating a set of potential query/answer pairs responsive to the text query. An answer identification routine compares the word phrases present in the user query with word phrases generated from each of the set of potential query/answer pairs to determine a best answer to the text query.

Claims

1. A query recognition system operating on a server coupled to the Internet comprising:

a natural language routine configured to generate a first linguistic analysis of words contained in a text query, including an identification of any word phrases present in said text query using one or both of grammars or dictionaries appropriate for a context of said text query;

a query formulation routine adapted to convert said words and said word phrases into a structured, combined query suitable for locating matching question/answer pairs for said text query;

wherein said first linguistic analysis is used to identify a candidate set of potential matching question/answer pairs for said text query; and

an answer identification routine configured for comparing said first linguistic analysis with one or more second linguistic analyses of words contained in said matching question/answer pairs so as to determine a best answer to said text query from the matching question/answer pairs, the best answer being determined from the question/answer pair that best matches the text query;

wherein the best answer is communicated over the Internet in response to said text query in real time.

2. The system of claim 1, wherein said query formulation routine generates a first level query using said words, and further customizes said first level query using natural language processing to generate the structured, combined query.

3. The system of claim 2, wherein said natural language routine operates to generate said word phrases during a time when said query formulation routine generates said first level query.

4. The system of claim 1, wherein the context of said text query relates to a computer program environment experienced by a user providing the text query, and is used for restricting said structured, combined query to a selected set of records in a question/answer pair database and/or an indexed database of question/answer pairs.

5. The system of claim 1, further including an Internet based server for executing said natural language routine, said query formulation routine, and said answer identification routine.

6. The system of claim 1, wherein said first linguistic analysis and said second analysis includes an operation for determining noun-phrases.

7. The system of claim 1, wherein said answer identification routine compares noun-phrases of said candidate set of potential matches with noun-phrases of said text query to determine said best answer.

8. The system of claim 1, wherein said structured, combined query is a full text query containing SQL search predicates.

9. The system of claim 1, wherein said potential matching question/answer pairs are retrieved from a relational database with indices that are updated asynchronously to reduce retrieval latency.

10. An Internet based query recognition and response system comprising:

a search engine for receiving a user query in text form;

a natural language routine configured to generate a first linguistic analysis of words contained in said user query, including an identification of first word phrases present in said user query, the natural language routine employing one of both of appropriate grammars or dictionaries based on a context of said user query;

wherein said first word phrases are extracted and stored in ranked order in an array;

a query formulation routine adapted to convert said first word phrases along with a list of all words in said text query into a structured, combined query including search predicates for locating a set of potential query/answer pairs responsive to said text query;

wherein second word phrases are extracted from said set of potential query/answer pairs and stored in ranked order in a second array; and

an answer identification routine configured for comparing said first word phrases with second word phrases to identify a responsive answer to said text query, the responsive answer being based on one of said set of potential query/answer pairs that best matches the user query.

11. The system of claim 10, wherein the context of said user query relates to a computer program environment experienced by a user providing the user query, and is used for restricting said structured query to a selected set of records in a question/answer pair database and/or an indexed database of question/answer pairs.

12. The system of claim 10, wherein said structured, combined query is a full text query containing SQL search predicates.

13. The system of claim 10, wherein said potential matching question/answer pairs are retrieved from a relational database with indices that are updated asynchronously to reduce retrieval latency.

14. The system of claim 10, wherein said first linguistic analysis and said second analysis includes an operation for determining noun-phrases.

15. The system of claim 10, wherein said answer identification routine compares noun-phrases of said candidate set of potential matches with noun-phrases of said text query to determine said best answer.

16. The system of claim 10, wherein the user query is submitted by a user using an application communicating with said search engine using an Internet-based communications protocol.

17. The system of claim 10, wherein said query formulation routine generates a first level query using said words, and further customizes said first level query using natural language processing to generate the structured, combined query.

18. The system of claim 17, wherein said natural language routine operates to generate said word phrases during a time when said query formulation routine generates said first level query.