Systems and methods for advanced grammar checking

• US 10,387,565 B2
• Filed: 08/29/2016
• Issued: 08/20/2019
• Est. Priority Date: 05/13/2010
• Status: Active Grant

• Current Assignee: Grammarly, Inc.
• Original Assignee: Grammarly, Inc.
• Inventors: Hoover, Bradley, Lytvyn, Maksym, Shevchenko, Oleksiy
• Primary Examiner(s): Azad, Abul K
• Application Number:
  US15/249,986
• Publication Number:
  US 20160371248A1
• Time in Patent Office: 1,086 Days
• Field of Search: 704 9, 704235, 704257, 704270, 7042701, 704275
• CPC Class Codes: G06F 40/205 Parsing
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G06F 40/253 Grammatical analysis; Style critique
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking

Abstract

In embodiments of the present invention improved capabilities are described for methods and systems of grammar checking comprising a grammar checking facility and a plurality of human proofreaders in a crowd-source population for the detection and correction of grammatical errors in text as received from a computing device, wherein possible detected errors are categorized as high confidence or low confidence possible errors, and the high confidence errors are distributed to the human proofreaders.

Claims

1. A method of grammar checking, the method comprising:

receiving text through a communications network at a first computing device-based grammar checking facility from a second computing device, wherein the grammar checking facility automatically detects a plurality of possible grammatical errors in the text through software algorithms, and categorizes each of the detected plurality of possible grammatical errors as one of at least a high confidence possible error and a low confidence possible error, wherein a high confidence possible error is considered an identified error and a low confidence possible error is one the grammar checking facility can not determine exists with high enough certainty;

providing, automatically from the first computing device-based grammar checking facility through the communications network to each of a plurality of human proofreader computing devices of a plurality of human proofreaders in a crowd-source population, at least one corresponding excerpt of the text containing a possible grammatical error categorized as a low confidence possible error as a result of the automated detection of the grammar checking facility so that an augmented error detection process is collectively performed by the plurality of human proofreaders in parallel for those grammatical errors categorized as low confidence possible errors to determine whether an actual error exists, wherein an error determined by a human proofreader to be an actual error is either corrected by the human proofreader to produce a correction result or is sent back to the grammar checking facility for automatic correction or for distribution as an identified error to at least one human proofreader, wherein the grammar checking facility may automatically correct at least some of the identified errors to produce corrected text portions;

managing, by the grammar checking facility, the automatic distribution of at least some of the identified errors to the plurality of human proofreaders, so that a human augmented grammar correction is collectively performed by the plurality of human proofreaders in parallel to produce correction results, wherein additional grammatical errors not found in the grammar checking facility automatic detection of the plurality of possible grammatical errors are detected by the plurality of human proofreaders, wherein the plurality of human proofreaders correct the additional grammatical errors and the distributed identified grammatical errors, and wherein each of the plurality of human proofreader computing devices is separate from the first computing device and the second computing device;

receiving, by the grammar checking facility, the correction results of the plurality of human proofreaders and automatically incorporating the correction results with any automatically corrected text portions of the grammar checking facility to contribute to an at least one corrected version of the received text; and

sending the at least one corrected version back to the second computing device.

2. The method of claim 1, wherein the second computing device interfaces with the grammar checking facility through an application programming interface.

3. The method of claim 2, wherein the second computing device is an enterprise server and the enterprise server is hosting a web accessible service of the grammar checking facility through the application programming interface.

4. The method of claim 2, wherein the application programming interface is associated with at least one of a third-party application, a word processor, a browser, a search engine, an email client, an SMS application, a blogging application, a micro-blogging application, a content management facility, a publishing facility, and an advertisement facility.

5. The method of claim 1, wherein the at least one corrected version includes a plurality of corrected excerpts integrated back into the original received text and wherein the second computing device receives the entire text with mistakes corrected.

6. The method of claim 1, wherein the at least one corrected version is a redline version.

7. The method of claim 1, wherein at least one of the plurality of human proofreader computing devices is a mobile computing device.

8. A system of grammar checking, the system comprising:

a first computing device-based grammar checking facility integrated into a computing environment to analyze text generated by a remote second computing device for grammatical errors, wherein the second computing device transmits the text to the grammar checking facility via a communications network, and the grammar checking facility automatically detects a plurality of possible grammatical errors in the text through software algorithms, categorizes each of the detected plurality of possible errors as one of at least a high confidence possible error and a low confidence possible error, wherein a high confidence possible error is considered an identified error, and a low confidence possible error is one the grammar checking facility can not determine exists with high enough certainty, and automatically sends, to each of a plurality of human proofreader computing devices of a plurality of human proofreaders in a crowd-source population, at least one corresponding excerpt of the text containing at least one of the plurality of possible grammatical errors categorized as a low confidence possible error so that an augmented error detection process collectively performed by the plurality of human proofreaders in parallel for those grammatical errors categorized as low confidence possible errors to determine whether an actual error exists, wherein an error determined by a human proofreader to be an actual error is either corrected by the human proofreader to produce a correction result or is sent back to the grammar checking facility for automatic correction or for distribution as an identified error to at least one human proofreader, wherein the grammar checking facility may automatically correct at least some of the some of the identified errors to produce corrected text portions, wherein the grammar checking facility manages the automatic distribution of at least some of the identified errors to the plurality of human proofreaders, so that a human augmented grammar correction is collectively performed by the plurality of human proofreaders in parallel to produce correction results, wherein additional grammatical errors not found in the grammar checking facility automatic detection of the plurality of possible grammatical errors are detected by the plurality of human proofreaders, wherein the plurality of human proofreaders correct the additional grammatical errors and the distributed identified grammatical errors, and wherein each of the plurality of human proofreader computing devices is separate from the first computing device and the second computing device,

wherein the correction results of the human augmented grammar checking by the plurality of human proofreaders are received by the grammar checking facility and are automatically incorporated with any automatically corrected text portions of the grammar checking facility to contribute to at least one corrected version of the received text, and

wherein the at least one corrected version is sent back to the second computing device.

9. The system of claim 8, wherein the second computing device is an enterprise server.

10. The system of claim 9, wherein the enterprise server is hosting a web accessible service of the grammar checking facility, wherein the second computing device transmits the text to the grammar checking facility through an application programming interface.

11. The system of claim 10, wherein the application programming interface is associated with at least one of a third-party application, a word processor, a browser, a search engine, an email client, an SMS application, a blogging application, a micro-blogging application, a content management facility, a publishing facility, and an advertisement facility.

12. A method of grammar checking, the method comprising:

receiving text through a communications network at a first computing device-based grammar checking facility from a second computing device, wherein the grammar checking facility automatically detects a plurality of possible grammatical errors in the text through software algorithms, categorizes each grammatical error of the detected plurality of possible grammatical errors as one of at least a high confidence possible error and a low confidence possible error, wherein the high confidence possible error is considered an identified error and a low confidence possible error is one the grammar checking facility can not determine exists with high enough certainty;

automatically transmitting each of the plurality of possible grammatical errors categorized as low confidence errors from the grammar checking facility through the communications network to a plurality of human proofreader computing devices of a plurality of human proofreaders in a crowd-source population, where those grammatical errors categorized as low confidence errors are distributed in parallel amongst the plurality of human proofreader computer devices, wherein each of the plurality of possible grammatical errors categorized as low confidence errors is transmitted to at least one of the plurality of human proofreader computing devices for error identification where the grammar checking facility could not determine if the error exists with high enough certainty, wherein at least one of the possible grammatical errors categorized as low confidence errors is transmitted to at least two human proofreaders of the plurality of human proofreaders for assessment, wherein one of the at least two human proofreaders performs error identification and one of the at least two human proofreaders performs error correction, wherein additional grammatical errors not found in the grammar checking facility automatic detection of the plurality of possible grammatical errors are detected by the plurality of human proofreaders, and wherein the additional grammatical errors and the identified errors are corrected by at least one of the grammar checking facility to produce corrected text portions and at least one of the plurality of human proofreaders to produce correction results, and wherein each of the plurality of human proofreader computing devices is separate from the first computing device and the second computing device;

receiving, by the grammar checking facility, any correction results of the plurality of human proofreaders, and incorporating any correction results of the plurality of human proofreaders with any corrected text portions of the grammar checking facility to contribute to a corrected version of the text; and

transmitting the assembled corrected text to the second computing device.

13. The method of claim 12, further comprising detecting, by the plurality of human proofreaders, additional grammatical errors not found in the grammar checking facility automatic detection of the plurality of possible grammatical errors, and correcting, by the plurality of human proofreaders, the additional grammatical errors in addition to correcting the plurality of grammatical errors categorized as low confidence possible errors detected through the grammar checking facility.

14. The method of claim 12, wherein at least one of the plurality of human proofreader computing devices is a mobile computing device.

15. The method of claim 12, wherein a user of the second computing device previously entered the received text into the second computing device.

16. The method of claim 15, wherein the user of the second computing device entered the received text through at least one of directly typing, uploading, and a copy-and-paste operation into an application on the second computing device.

17. The method of claim 15, wherein the user entered the received text though an application program interface (API) associated with the grammar checking facility.

18. The method of claim 12, further comprising grammar checking the assembled corrected text prior to transmitting the assembled corrected text to the second computing device.

19. The method of claim 18, wherein the grammar checking of the assembled corrected text is provided at least in part by the grammar checking facility.

20. The method of claim 18, wherein the grammar checking of the assembled corrected text is provided at least in part by the plurality of human proofreaders.