Web-based proofing and usage guidance

• US 7,991,609 B2
• Filed: 02/28/2007
• Issued: 08/02/2011
• Est. Priority Date: 02/28/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Brockett, Chris, Hon, Hsiao-Wen, Kacmarcik, Gary, Gao, Jianfeng, Klementiev, Alexandre, Vanderwende, Lucy, Zhou, Ming, Dolan, William, Gamon, Michael
• Primary Examiner(s): Hudspeth; David R
• Assistant Examiner(s): Spooner; Lamont M
• Application Number:
  US11/713,073
• Publication Number:
  US 20080208567A1
• Time in Patent Office: 1,616 Days
• Field of Search: 704/1, 704/9, 704/10, 715/255, 715/256, 715/257
• US Class Current: 704/9
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G06F 40/253 Grammatical analysis; Style critique

Abstract

A system is disclosed for checking grammar and usage using a flexible portfolio of different mechanisms, and automatically providing a variety of different examples of standard usage, selected from analogous Web content. The system can be used for checking the grammar and usage in any application that involves natural language text, such as word processing, email, and presentation applications. The grammar and usage can be evaluated using several complementary evaluation modules, which may include one based on a trained classifier, one based on regular expressions, and one based on comparative searches of the Web or a local corpus. The evaluation modules can provide a set of suggested alternative segments with corrected grammar and usage. A followup, screened Web search based on the alternative segments, in context, may provide several different in-context examples of proper grammar and usage that the user can consider and select from.

Claims

1. A method, implemented at least in part by a computing system that has a processor, comprising:

evaluating, with the processor, whether one or more strings of text include non-standard usage by;

performing at least one search restricted to a first set of one or more data stores indicated to have relatively more reliable language usage in a language of the strings of text; and

performing at least one search restricted to a second set of one or more data stores indicated to have relatively less reliable language usage in the language of the strings of text;

wherein a relative proportion of matches from search results of the second set compared with the first set is used as an indicator of non-standard usage;

providing an indication that a string of text includes non-standard usage, with one or more suggested alternative strings of text; and

providing, with the processor, one or more examples of usage of the one or more suggested alternative strings of text.

2. The method of claim 1, further comprising using one or more evaluation modules to evaluate whether the one or more strings of text include non-standard usage, and using the one or more evaluation modules to provide the one or more suggested alternative strings of text.

3. The method of claim 2, wherein one of the evaluation modules extracts lexical features from the strings of text and evaluates the lexical features with a classifier.

4. The method of claim 3, wherein the classifier further provides at least one of the suggested alternative strings of text, based on at least one lexical feature found to be a relatively close match for at least one string of text with non-standard usage.

5. The method of claim 2, wherein one of the evaluation modules compares the strings of text with a set of regular expressions.

6. The method of claim 2, wherein one of the evaluation modules performs the at least one search through the first and second sets of data stores and compares the one or more strings of text with a the data stores, and evaluates whether the one or more strings of text conform with standard usage in the data stores.

7. The method of claim 6, wherein evaluating whether the one or more strings of text conform with standard usage in the data stores comprises evaluating the prevalence of either or both of exact matches or fuzzy matches of the one or more strings of text in the data stores.

8. The method of claim 6, wherein the data stores comprise Web content accessible to a Web search, and wherein comparing the strings of text with the corpus comprises comparing the strings of text with results from one or more Web searches for queries that include the strings of text or are analogous to the strings of text.

9. The method of claim 8 wherein performing at least one search restricted to the first set of data stores includes performing at least one Web search wherein the first set of one or more data stores comprises a first set of one or more domains and wherein performing at least one search restricted to the second set of data stores includes performing at least one Web search wherein the second set of one or more data stores comprises a second set of one or more domains.

10. The method of claim 2, further comprising detecting a subject language in which the strings of text are written, and directing the evaluation modules to evaluate the strings of text with reference to the subject language.

11. The method of claim 2, wherein the evaluation modules are accessed through an application protocol interface, and the method further comprises adding one or more additional evaluation modules conforming to the application protocol interface, and using at least one of the added evaluation modules to evaluate the strings of text.

12. The method of claim 2, wherein the one or more evaluation modules generate a plurality of suggested alternative strings of text, and wherein the method further comprises filtering out suggested alternative strings of text indicated to be unlikely, unhelpful, or duplicative.

13. The method of claim 1, wherein evaluating whether the one or more strings of text include non-standard usage comprises targeting non-standard usage involving one or more particular grammatical elements.

14. The method of claim 1, further comprising screening a plurality of candidate alternative strings of text before providing the suggested alternative strings of text, and ranking the suggested alternative strings of text in descending order of a language model score.

15. The method of claim 1, wherein the one or more examples of usage of the one or more suggested alternative strings of text are provided in response to an option for viewing the examples being selected, after the one or more suggested alternative strings of text are provided.

16. The method of claim 1, wherein providing the one or more examples of usage of the one or more alternative strings of text comprises performing a search for the alternative strings of text in at least one of the Web or a discrete corpus, and providing results from the search.

17. The method of claim 16, wherein performing the search for the alternative strings of text comprises using a query that includes one of the strings of text and one or more additional words from a sentence from which the string of text is derived.

18. The method of claim 1, further comprising providing the examples of usage of the alternative strings of text in an order ranked by criteria that include either or both of:

similarity of context, and brevity.