Machine assisted query formulation

• US 8,190,627 B2
• Filed: 06/28/2007
• Issued: 05/29/2012
• Est. Priority Date: 06/28/2007
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Platt, John C., Flake, Gary W., Naam, Ramez, Gupta, Anoop, Hurst-Hiller, Oliver, Griffin, Trenholme J., Goodman, Joshua T.
• Primary Examiner(s): Jalil, Neveen Abel
• Assistant Examiner(s): KINSAUL, DANIEL W
• Application Number:
  US11/770,458
• Publication Number:
  US 20090006343A1
• Time in Patent Office: 1,797 Days
• Field of Search: None
• US Class Current: 707/767
• CPC Class Codes: G06F 16/242 Query formulation

Abstract

Architecture for completing search queries by using artificial intelligence based schemes to infer search intentions of users. Partial queries are completed dynamically in real time. Additionally, search aliasing can also be employed. Custom tuning can be performed based on at least query inputs in the form of text, graffiti, images, handwriting, voice, audio, and video signals. Natural language processing occurs, along with handwriting recognition and slang recognition. The system includes a classifier that receives a partial query as input, accesses a query database based on contents of the query input, and infers an intended search goal from query information stored on the query database. A query formulation engine receives search information associated with the intended search goal and generates a completed formal query for execution.

Claims

1. A computer-implemented system that facilitates Internet-based interactions for machine assisted query formulation, comprising:

a processor;

a context component executed by the processor that generates context data of a user interacting in context of a word processing application or a programming application, wherein the user interacting comprises initiating a search from within the word processing application or the programming application;

a syntactical analyzer component executed by the processor that receives a first partial query input including at least a stylus-based stroke query input and performs a syntactical analysis on the first partial query input to generate syntactically analyzed partial query input;

a stylus-based stroke recognition component that receives the stylus-based stroke query input and performs stylus-based stroke recognition interpretation and analysis on the stylus-based stroke query input to generate stylus-based stroke data;

a support vector machine (SVM) classifier component that receives the syntactically analyzed partial query input from the syntactical analyzer component and the stylus-based stroke data from the stylus-based stroke recognition component, infers an intended search goal based on the context data, and augments the first partial query input including the stylus-based stroke-type query input to reflect the intended goal of the search;

a voice recognition component that receives a second partial query input including at least utterances, spoken natural language terms, and phrases, and performs an analysis and processing on the second partial query input based on the context data;

a recognition component that performs an analysis and processing on the first partial query input and the second partial query input based at least on different types of sequential partial query inputs and in the context data; and

a query formulation component executed by the processor that generates a formal query based on the first partial query input and the second partial query input based on the word processing application or the programming application of the context data.

2. The system of claim 1, further comprising a query database that includes indexed data from which the intended search goal is inferred.

3. The system of claim 1, wherein the first partial query input comprises a combination of at least two or more of the following:

textual data, graffiti, handwriting, images, audio, or voice signals.

4. The system of claim 1, wherein the first partial query input further includes at least one of graffiti, handwriting, audio, or images.

5. The system of claim 1, further comprising a modeling component executed by the processor for generating a voice recognition model based on at least one of the context data, location, terms, phrases, or speech characteristics.

6. The system of claim 1, wherein the first partial query input and the second partial query input are in the form of programming code having terms and characters, the query formulation component generating the formal query based on the terms and characters.

7. The system of claim 1, wherein the formal query includes a word that is different than a completed word of the first partial query input.

8. The system of claim 1, further comprising a machine learning and reasoning component executed by the processor that employs at least one of a probabilistic or a statistical-based analysis to prognose or infer an action that the user desires to be automatically performed.

9. The system of claim 1, wherein the context component executed by the processor that determines context data associated with a software environment in which the user is currently active and, utilizes the context data and user identity data to generate the formal query.

10. The system of claim 1, further comprising a geolocation component executed by the processor that computes geographic location data of a user, and that is used to infer a formal query.

11. The system of claim 1, wherein the classifier infers the intended search goal from data of multiple signal input types used to enter the first partial query input.

12. The system of claim 1, further comprising a machine learning and reasoning component executed by the processor that learns and reasons about user interaction with the first partial query input and the formal query.

13. A computer implemented method of completing search input for inferring search intentions of users executing on a processing unit, comprising:

determining a user context of a user logged on a context of a browser;

receiving a first partial query input via a plurality of inputs, the first partial query input includes at least a stylus-based stroke query input;

accessing a datasource in response to receiving the first partial query input;

performing recognition and analysis of the stylus-based stroke-type query input to generate stylus-based stroke data;

inferring an intended query from the first partial query input by syntactically analyzing the first partial query input to generate a syntactically analyzed partial query input;

processing the syntactically analyzed partial query input and the stylus-based stroke data with a support vector machine, and augmenting the first partial query input including the at least stylus-based stroke-type query input to reflect an intended search goal, the first partial query input being related to websites associated with the browser of the user context;

analyzing a structure of the ways words and symbols are used in the first partial query input, and comparing common language usage and more specific syntax properties for resolving a formal query;

receiving a second partial query input via the plurality of inputs, the second partial query input includes at least utterances, spoken natural language terms, and phrases, the second partial query input related to information associated with the browser of the user context;

performing an analysis and processing on the first partial query input and the second partial query input based at least on different types of sequential partial query inputs in context of the browser, and the user context associated with the browser, to facilitate formulating the formal query; and

formulating the formal query based on the user context, in relationship with the first partial and the second partial query information.

14. The method of claim 13, further comprises performing recognition processing on input signals used to create the first partial query input.

15. The method of claim 13, further comprises presenting multimedia content as part of the formal query.

16. The method of claim 13, further comprises monitoring user interaction with the formal query and, learning and reasoning about the user interaction to refine a future partial query input.

17. The method of claim 13, further comprising setting a predetermined threshold value for providing a qualitative measure of a training process.

18. The method of claim 13, further comprising processing multiple input types to generate the first partial query input, and formulating the formal query based on inferences made on corresponding sets of query information for the multiple input types.

19. The method of claim 13, further comprising assigning points to a query completion word based on length of the word.

20. A computer-readable storage media storing computer-readable instructions for machine assisted query formulation, executed on a processing unit;

the computer-readable storage media comprising;

a query input processing component configured to receive input from a plurality of input devices and to generate a first partial query input, the first partial query input includes at least a stylus-based stroke-type query input;

a stylus-based stroke recognition component configured to interpret and to analyze the stylus-based stroke-type input query to generate stylus-based stroke data;

a partial query formulation component configured to determine user context of a user associated with a context of a browser, to perform a syntactical analysis of the first partial query input having a high probability of being related to historical websites associated with the browser, and to infer an intended search goal from the syntactically analyzed partial query input and from the stylus-based stroke data with a support vector machine;

a voice recognition component configured to receive input from the plurality of input devices and to generate a second partial query input, the second partial query input includes at least utterances, spoken natural language terms, and phrases, and performs an analysis and processing on the second partial query input;

a recognition component configured to perform an analysis and to process the first partial query input and the second partial query input based at least on different types of sequential partial query inputs, and based on the user context and the first partial query input being related to the historical websites associated with the browser; and

a formal query component configured to generate a formal query based on the inferred intended search goal of the user context, the first partial query input and the second partial query input.