IP Library Granted Patent US 12,566,805
Granted Patent B1
US 12,566,805 · App. 18/542,224 · Granted Mar 3, 2026

Generating context-based spell corrections of entity names

Inventors: Lawrence J. Brunsman (Mountain View, CA); Matthieu Devin (Los Altos Hills, CA); Uri N. Lerner (Los Altos, CA); Simon Tong (Mountain View, CA)
Assignee: GOOGLE LLC
G06F16/951G06F16/24G06F16/3338G06F40/232G06F40/295G06F40/247
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,566,805
App. No.
18/542,224
Granted
Mar 3, 2026
Kind
B1
Abstract

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for correcting entity names. One method includes receiving texts and deriving a plurality of name-context pairs from the texts. The method further includes calculating a context consistency measure for each name-context pair and storing context-entity name data representing the name-context pairs. Another method includes identifying an entity name and one or more context terms from a query and generating candidate names for the entity name. The method further includes determining a score for each of the candidate names, selecting a number of top scoring candidate names, and using the selected candidate names to respond to the query.

Claims (42)

1 . A method implemented by one or more processors, the method comprising:

providing a search query interface;

receiving a search query, through the search query interface, that requests information about a particular entity, the search query comprising (i) a name portion comprising one or more misspelled words representing an entity name, and (ii) a context portion comprising one or more context words descriptive of the particular entity, wherein the search query is received as a single query string that includes both the name portion and the context portion, and the entity name corresponds to each of a plurality of candidate entities that are potentially responsive to the search query;

processing, responsive to receiving the search query through the search query interface, the search query to identify the name portion and the context portion;

identifying a particular candidate entity of the plurality of candidate entities as responsive to the request for information about the particular entity, wherein identifying the particular candidate entity as responsive to the request for information about the particular entity is based on co-occurrences of the one or more context words in the context portion of the search query with the particular candidate entity, and wherein the particular candidate entity corresponds to a corrected spelling of the one or more misspelled words representing the entity name; and

in response to identifying the particular candidate entity of the plurality of candidate entities as responsive to the request for information about the particular entity, automatically correcting the search query by replacing the one or more misspelled words with the corrected spelling, and providing a search results page that includes search results generated based on the automatically corrected search query that replaces the one or more misspelled words with the corrected spelling.

2 . The method according to claim 1 , wherein the search results page identifies the particular candidate entity of the plurality of candidate entities that corresponds to the corrected spelling of the entity name.

3 . The method according to claim 2 , wherein the search results page identifies at least one of the one or more context words.

4 . The method according to claim 2 , wherein the search results page identifies the name portion of the search query separately from the particular candidate entity.

5 . The method according to claim 1 , wherein the search results page includes an image of the particular candidate entity of the plurality of candidate entities that corresponds to the corrected spelling of the entity name.

6 . The method according to claim 1 , further comprising:

receiving a selection of a selectable control; and

based on the selection of the selectable control, providing another search results page that includes additional information related to one or more other candidate entities.

7 . The method according to claim 1 , wherein the co-occurrences are in a plurality of documents indexed by a search engine.

8 . A computer program product comprising one or more non-transitory computer-readable storage media having program instructions collectively stored on the one or more computer-readable storage media, the program instructions executable to:

provide a search query interface;

receive a search query, through the search query interface, that requests information about a particular entity, the search query comprising (1) a name portion comprising one or more misspelled words representing an entity name, and (ii) a context portion comprising one or more context words descriptive of the particular entity, wherein the search query is received as a single query string that includes both the name portion and the context portion, and the entity name corresponds to each of a plurality of candidate entities that are potentially responsive to the search query;

process, responsive to receiving the search query through the search query interface, the search query to identify the name portion and the context portion;

identify a particular candidate entity of the plurality of candidate entities as responsive to the request for information about the particular entity, wherein identifying the particular candidate entity as responsive to the request for information about the particular entity is based on co-occurrences of the one or more context words in the context portion of the search query with the particular candidate entity, and wherein the particular candidate entity corresponds to a corrected spelling of the one or more misspelled words representing the entity name; and

in response to identifying the particular candidate entity of the plurality of candidate entities as responsive to the request for information about the particular entity, automatically correct the search query by replacing the one or more misspelled words with the corrected spelling, and provide a search results page that includes search results generated based on the automatically corrected search query that replaces the one or more misspelled words with the corrected spelling.

9 . The computer program product according to claim 8 , wherein the search results page identifies the particular candidate entity of the plurality of candidate entities that corresponds to the corrected spelling of the entity name.

10 . The computer program product according to claim 9 , wherein the search results page identifies at least one of the one or more context words.

11 . The computer program product according to claim 9 , wherein the search results page identifies the name portion of the search query separately from the particular candidate entity.

12 . The computer program product according to claim 8 , wherein the search results page includes an image of the particular candidate entity of the plurality of candidate entities that corresponds to the corrected spelling of the entity name.

13 . The computer program product according to claim 8 , wherein the program instructions are further executable to:

receive a selection of a selectable control; and

based on the selection of the selectable control, provide another search results page that includes additional information related to one or more other candidate entities.

14 . The computer program product according to claim 8 , wherein the co-occurrences are in a plurality of documents indexed by a search engine.

15 . A system comprising:

a processor, a computer-readable memory, one or more non-transitory computer-readable storage media, and program instructions collectively stored on the one or more computer-readable storage media, the program instructions executable to:

provide a search query interface;

receive a search query, through the search query interface, that requests information about a particular entity, the search query comprising (1) a name portion comprising one or more misspelled words representing an entity name, and (ii) a context portion comprising one or more context words descriptive of the particular entity, wherein the search query is received as a single query string that includes both the name portion and the context portion, and the entity name corresponds to each of a plurality of candidate entities that are potentially responsive to the search query;

process, responsive to receiving the search query through the search query interface, the search query to identify the name portion and the context portion;

identify a particular candidate entity of the plurality of candidate entities as responsive to the request for information about the particular entity, wherein identifying the particular candidate entity as responsive to the request for information about the particular entity is based on co-occurrences of the one or more context words in the context portion of the search query with the particular candidate entity, and wherein the particular candidate entity corresponds to a corrected spelling of the one or more misspelled words representing the entity name; and

in response to identifying the particular candidate entity of the plurality of candidate entities as responsive to the request for information about the particular entity, automatically correct the search query by replacing the one or more misspelled words with the corrected spelling, and provide a search results page that includes search results generated based on the automatically corrected search query that replaces the one or more misspelled words with the corrected spelling.

16 . The system according to claim 15 , wherein the search results page identifies the particular candidate entity of the plurality of candidate entities that corresponds to the corrected spelling of the entity name.

17 . The system according to claim 16 , wherein the search results page identifies at least one of the one or more context words.

18 . The system according to claim 16 , wherein the search results page identifies the name portion of the search query separately from the particular candidate entity.

19 . The system according to claim 15 , wherein the search results page includes an image of the particular candidate entity of the plurality of candidate entities that corresponds to the corrected spelling of the entity name.

20 . The system according to claim 15 , wherein the program instructions are further executable to:

receive a selection of a selectable control; and

based on the selection of the selectable control, provide another search results page that includes additional information related to one or more other candidate entities.

Assignments (2)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 19, 2023
From: BRUNSMAN, LAWRENCE J.; DEVIN, MATTHIEU; LERNER, URI N.; TONG, SIMON
To: GOOGLE INC.
Reel/Frame 065914/0347 →
CHANGE OF NAME Recorded Dec 19, 2023
From: GOOGLE INC.
To: GOOGLE LLC
Reel/Frame 066076/0877 →
Continuity (4)
Continuation 16227086 · Dec 20, 2018
Continuation 14616915 · Feb 9, 2015
Continuation 13071392 · Mar 24, 2011
Provisional Application 61317622 · Mar 25, 2010
References Cited (55)
US 4499553A · Dickinson · 1985 [cited by examiner]
US 5542090A · Henderson · 1996 [cited by applicant]
US 6144958A · Ortega et al. · 2000 [cited by applicant]
US 6167397A · Jacobson · 2000 [cited by applicant]
US 6236768B1 · Rhodes · 2001 [cited by applicant]
US 6271840B1 · Finseth et al. · 2001 [cited by applicant]
US 6446064B1 · Livowsky · 2002 [cited by applicant]
US 6601059B1 · Fries · 2003 [cited by applicant]
US 6742176B1 · Million · 2004 [cited by applicant]
US 7231393B1 · Harik · 2007 [cited by applicant]
US 7296019B1 · Chandrasekar · 2007 [cited by applicant]
US 7383258B2 · Harik et al. · 2008 [cited by applicant]
US 7444324B2 · Ortega · 2008 [cited by examiner]
US 7603349B1 · Kraft et al. · 2009 [cited by applicant]
US 7818332B2 · Olds · 2010 [cited by applicant]
US 7958115B2 · Kraft · 2011 [cited by applicant]
US 8180760B1 · Carver · 2012 [cited by applicant]
US 8296257B1 · Chavira · 2012 [cited by applicant]
US 8392396B1 · Carver · 2013 [cited by applicant]
US 8402032B1 · Brunsman et al. · 2013 [cited by applicant]
US 8538984B1 · Mahabal et al. · 2013 [cited by applicant]
US 9002866B1 · Brunsman · 2015 [cited by applicant]
US 10162895B1 · Brunsman · 2018 [cited by applicant]
US 20020059219A1 · Neveitt · 2002 [cited by applicant]
US 20040236725A1 · Amitay et al. · 2004 [cited by applicant]
US 20050033641A1 · Jha et al. · 2005 [cited by applicant]
US 20050210017A1 · Cucerzan · 2005 [cited by applicant]
US 20050210383A1 · Cucerzan · 2005 [cited by examiner]
US 20060026204A1 · Campbell · 2006 [cited by applicant]
US 20060047636A1 · Mohania et al. · 2006 [cited by applicant]
US 20060200464A1 · Gideoni · 2006 [cited by applicant]
US 20060206454A1 · Forstall et al. · 2006 [cited by applicant]
US 20060280309A1 · Krishnaswamy · 2006 [cited by applicant]
US 20070038601A1 · Guha · 2007 [cited by applicant]
US 20070094234A1 · Wen · 2007 [cited by applicant]
US 20070174299A1 · Kuang et al. · 2007 [cited by applicant]
US 20070185777A1 · Pyle · 2007 [cited by applicant]
US 20070294615A1 · Sathe · 2007 [cited by applicant]
US 20080046405A1 · Olds et al. · 2008 [cited by applicant]
US 20080104071A1 · Pragada · 2008 [cited by examiner]
US 20080168053A1 · Garg · 2008 [cited by applicant]
US 20080183672A1 · Canon et al. · 2008 [cited by applicant]
US 20090070325A1 · Gabriel · 2009 [cited by applicant]
US 20090313100A1 · Ingleshwar · 2009 [cited by applicant]
US 20100180198A1 · Iakobashvili et al. · 2010 [cited by applicant]
US 20100180199A1 · Wu et al. · 2010 [cited by applicant]
US 20100293179A1 · Chaudhuri · 2010 [cited by applicant]
US 20110035403A1 · Ismalon · 2011 [cited by applicant]
US 20110106807A1 · Srihari et al. · 2011 [cited by applicant]
US 20110225155A1 · Roulland et al. · 2011 [cited by applicant]
US 20110282888A1 · Koperski · 2011 [cited by applicant]
US 20110307432A1 · Yao et al. · 2011 [cited by applicant]
Martins, et al. Spelling Correction for Search Engine Queries, J. L. Vicedo et al. (Eds.): EsTAL 2004, LNAI 3230, pp. 372-383, 2004. (Year: 2004). [cited by examiner]
Niu et al., Word Independent Context Pair Classification Model for Word Sense Disambiguation, Proceedings of the 9th Conference on Computational Natural Language Learning (CoNLL), pp. 33-39, Ann Arbor, dated Jun. 2005. [cited by applicant]
Niu et al., Weakly supervised learning for cross-document person name disambiguation supported by information extraction, ACL '04 Proceedings of the 42nd Annual Meeting on Association of Computational Linguistics, pp. 1… [cited by applicant]