WO2017116471A1 - Identification d'erreurs dans des données d'entrée provenant de sources multiples - Google Patents

Identification d'erreurs dans des données d'entrée provenant de sources multiples Download PDF

Info

Publication number
WO2017116471A1
WO2017116471A1 PCT/US2015/068296 US2015068296W WO2017116471A1 WO 2017116471 A1 WO2017116471 A1 WO 2017116471A1 US 2015068296 W US2015068296 W US 2015068296W WO 2017116471 A1 WO2017116471 A1 WO 2017116471A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
sources
error
information
errors
Prior art date
Application number
PCT/US2015/068296
Other languages
English (en)
Inventor
Joel Fogelson
Arden ASH
Original Assignee
Technicolor Usa, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Technicolor Usa, Inc. filed Critical Technicolor Usa, Inc.
Priority to PCT/US2015/068296 priority Critical patent/WO2017116471A1/fr
Publication of WO2017116471A1 publication Critical patent/WO2017116471A1/fr

Links

Classifications

    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
    • G09B19/00—Teaching not covered by other main groups of this subclass
    • G09B19/04—Speaking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/232—Orthographic correction, e.g. spell checking or vowelisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/253—Grammatical analysis; Style critique
    • G—PHYSICS
    • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
    • G09B19/00—Teaching not covered by other main groups of this subclass
    • G09B19/06—Foreign languages

Definitions

  • the present disclosure generally relates to analyzing user input data to identify errors and, more particularly, to aggregating user input data and analyzing the aggregated data to distinguish errors associated with human error from errors related to other causes.
  • HMIs human-machine interfaces
  • a user might find it more difficult to type on a small virtual keyboard, and as a result, might make more errors when typing on a smartphone. That is, the user might know how to spell a particular word, but may type the word incorrectly more often on a virtual keyboard than on a full-sized mechanical keyboard.
  • the user's spelling errors for the word on a virtual keyboard may not be the result of human error (e.g., not knowing how to spell the word), but may be the result of a different cause (e.g., difficulty typing on a virtual keyboard).
  • error information can be obtained based on user input from a plurality of sources.
  • the error information can include information of misspelled words, information of violations of grammar rules, etc.
  • the plurality of sources can include devices, and each device can include one of a plurality of different types of HMIs.
  • the HMIs can include a mechanical keyboard and a virtual keyboard.
  • the plurality of sources can include different types of computer programs.
  • different types of computer programs can include word processing programs, texting programs, email programs, etc.
  • An input error can be determined from the error information.
  • Association information based on determined association can be stored and used, for example, to improve existing computer functionality.
  • a computer function can be performed based on the association information.
  • the function can include, for example, a spell checking function, a grammar checking function, a user-quizzing function, etc.
  • FIG. 1 illustrates an example of a computer network 100 in which various embodiments may be implemented.
  • FIG. 2 illustrates an example of an apparatus in which various embodiments may be implemented.
  • FIG. 3 is a flowchart of an example of a method for distinguishing human errors from errors associated with other causes according to various embodiments.
  • FIG. 4 is a flowchart of an example of a method for determining words to include in a spelling quiz based on distinguishing human errors from HMI-related errors according to various embodiments.
  • FIG. 5 is a flowchart of an example of a method for modifying an auto-suggest function of a grammar checking application based on distinguishing human errors from errors associated with a type of computer program according to various embodiments.
  • HMI human-machine interface
  • some input errors may be related to human error and some errors may be related to the HMI itself.
  • some people find it difficult to type on virtual keyboards, e.g., keyboards displayed on the touchscreens of smartphones, tablet computers, etc.
  • virtual keyboards e.g., keyboards displayed on the touchscreens of smartphones, tablet computers, etc.
  • some users can make more spelling errors when typing on a virtual keyboard than when typing on a mechanical keyboard, e.g., a typical keyboard connected to a personal computer.
  • Other data entry systems such as the use of a stylus with a virtual keyboard can be utilized as well.
  • some users might find it easier to type on a virtual keyboard, and may make fewer spelling errors on a virtual keyboard than on a mechanical keyboard. These users may make fewer errors on a virtual keyboard because, for example, the small size of virtual keyboards makes it easier to find the next key to press, or because the virtual keyboard is close to the location that the text appears. In some cases people use techniques for a virtual keyboard using two fingers, which may be more efficient for the person than using a standard mechanical keyboard.
  • determining which errors are human errors and which errors related to a particular HMI can provide useful information. For example, in order to help a user improve at spelling, a computer program can analyze user input data to determine words that the user commonly misspells when typing. The program can then quiz the user on the words that the user commonly misspells. If the user types on multiple devices, e.g., smartphone, personal computer, tablet computer, etc., the program can collect and aggregate the user input data across all of the devices in order to obtain a larger database to analyze. Having a larger database of user input can improve the accuracy of the error analysis. However, because the aggregated information can include HMI-related errors, such an approach can be improved upon.
  • the spelling quiz program may determine that the user sometimes types the "opem” instead of “open,” and as a result, may determine that the user should be quizzed on the spelling of the word "open.”
  • the user may know how to spell the word "open,” and the error may occur only when the user types on a virtual keyboard because of the user's difficulty with that particular HMI. Therefore, quizzing the user on the word "open” would not be helpful.
  • distinguishing spelling errors associated with human error from HMI-related spelling errors can improve the spelling quiz program by allowing the HMI-related misspellings to be given a low priority or weight, or to be completely disregarded, which can help focus the quiz on words that the user actually has trouble spelling.
  • Error correction is another area that can benefit from distinguishing human errors from HMI-related errors.
  • spell-checking functionality on personal computers, smartphones, tablets, etc. can analyze a user's typing errors to learn which errors the user commonly makes, and can use the analysis to provide better suggestions and/or auto-corrections.
  • the machine learning improves as more user input data is collected and analyzed. Therefore, collecting and aggregating user input data from all of the user's devices (e.g., personal computer, smartphone, tablet computer, etc.) and analyzing the aggregated data can allow the spell-checker to make better suggestions and/or auto-corrections than if spell-checkers (generically called an analyzer program) in each device analyzed the user input data from that device only.
  • spell-checkers generically called an analyzer program
  • the knowledge of which errors are common to all devices (e.g., human errors) and which errors occur only on particular devices (e.g., errors related to HMI) can be used to improve the automatic suggestions and/or corrections. For example, based on this additional knowledge, spell-checking can generate the "line" pop-up suggestion when the user types "line” on the virtual keyboard, but not when the user types "line” on the personal computer.
  • spell-checking can generate the "line” pop-up suggestion when the user types "line” on the virtual keyboard, but not when the user types "line” on the personal computer.
  • some user input errors may be associated with a particular type of computer program.
  • a user may use better grammar when typing in word processing applications than when typing in texting applications. This might be because the user prefers not to strictly adhere to the rules of grammar when texting, but uses word processing programs to prepare formal documents that require correct grammar.
  • the user may always obey a particular grammatical rule (rule A). However, the user may disregard rule A when using a texting application.
  • a grammar-quizzer program may aggregate user input data from different computer applications, including texting and word processing applications, and analyze the aggregated data to determine the user's grammar errors. If the grammar-quizzer does not distinguish errors made across all applications (i.e., human errors) from errors made only in particular applications (i.e., application-related errors) the grammar-quizzer may determine that the user has trouble with rule A and rule B. However, if the grammar-quizzer distinguishes human errors from application-related errors, the grammar-quizzer may make the more accurate determination that the user knows rule A, but has trouble with rule B.
  • Aggregated user input data can be analyzed to distinguish human errors from errors associated with other causes as well. For example, some errors may be associated with a particular person or category of person with which the user is communicating, e.g., a user may use better grammar when emailing a work associate than when emailing a friend. Distinguishing human errors from errors associated with communicating with friends may allow a grammar-quizzer to avoid testing the user on rules the user already knows. In another example, some errors may be associated with a time of day, e.g., a user may tend to make more typing errors late at night, such as misspelling words the user knows how to spell.
  • Distinguishing human errors from errors associated with typing late at night may allow a spell-checker to give lower weight to errors made late at night, which may improve the accuracy of suggestions and/or auto-corrections. It should be appreciated that errors may be associated with many reasons other than human error.
  • the other causes of error may be specifically targeted in the analysis of aggregated data.
  • a spell-checker may specifically search for differences in errors in data collected from mechanical keyboard devices and errors in data collected from virtual keyboard devices, or using a stylus interface.
  • the analysis may include determining other causes of error.
  • statistical analysis may be applied to determine likely causes of error other than human error.
  • a grammar-checker might not initially know that some grammar errors are associated with a user's preference not to use strict grammar when texting, but the grammar-checker can discover this association through statistical analysis of the aggregated user input data.
  • the grammar-checker can distinguish between grammar errors associated with human error (i.e., errors made regardless of whether the user is using texting applications, word processing applications, etc.) from errors made only in texting applications.
  • a statistical analysis may determine that the user disregards certain grammatical rules when typing emails using a personal email account and typing text messages using any texting application.
  • a grammar-checker may group these types of communications together in a "Suspect Grammar" category, and the analysis of the aggregated data can then distinguish human error from errors associated with the Suspect Grammar category of applications.
  • FIG. 1 illustrates an example of a computer network 100 in which various embodiments may be implemented.
  • Computer network 100 can include a storage server 101, a storage device 103 storing a database 105, devices 107A-107J, a local area network (LAN) 109, and an external network 111.
  • Storage device 103 can be, for example, a hard disk drive (HDD), a solid-state drive (SSD), etc., that stores database 105.
  • Database 105 as a computer implemented program, for some embodiments, can include, for example, aggregated user input data.
  • Storage server 101 can be, for example, a computer server connected to storage device 103.
  • Storage server 101 can include hardware, such as one or more processors and memory, software, and/or firmware to perform various functions, including aggregating user input data from multiple devices (such as devices 107A-107J), applications (such as word processing applications, email applications, texting applications), etc., analyzing the aggregated user input data to distinguish human errors from errors associated with other causes, and utilizing the results of the distinction of human errors according to the methods described in more detail below.
  • Devices 107A-107J can be, for example, desktop computers, laptop computers, tablet computers, mobile devices, video game systems, set top boxes, televisions, audio-video equipment, etc.
  • Devices 107A-107J can be operated by the various users, for example, and a single user can operate multiple devices.
  • devices 107A, 107J, and 1071 can be, respectively, a personal computer, a smartphone, and a tablet computer operated by a single user.
  • Devices 107A-107J can connect to storage server 101 through external network 111.
  • External network 111 can be, for example, the Internet.
  • Devices 107 A and 107B can connect to storage server 101 through LAN 109.
  • Various computer programs running on devices 107A-107J can connect to storage server 101 through external network 111, and various user input data can be transmitted to database 105 for aggregation and analysis in accordance with various embodiments described herein.
  • FIG. 2 illustrates an example of an apparatus in which various embodiments may be implemented.
  • FIG. 2 is a block diagram of an apparatus 200 for implementing various techniques described herein.
  • Apparatus 200 may be implemented, for example, as a general-purpose computing platform.
  • storage server 101, storage device 103, devices 107 A-J may be examples of apparatus 200.
  • Apparatus 200 can include a processor 210 for executing the computer- executable programs that perform various techniques described herein.
  • the programs may be stored in a memory 220, which may also store aggregated user input data.
  • a bus 230 can connect processor 210 and memory 220 to each other and to other components of apparatus 200.
  • apparatus 200 may include multiple processors or processors with multiple processing cores, which may execute various parts of programs in parallel.
  • a mass storage device 240 can be connected to bus 230 via a disk controller 250.
  • Mass storage device 240 may contain aggregated user input data, as well as an operating system, other programs, other data, etc.
  • Disk controller 250 may operate according to Serial Advanced Technology Advancement (SATA), Small Computer System Interface (SCSI), or other standards, and may provide connection to multiple mass storage devices.
  • SATA Serial Advanced Technology Advancement
  • SCSI Small Computer System Interface
  • a video display 260 can be connected to bus 230 via a video controller 270.
  • Video controller 270 may provide its own memory and graphics-processing capability for providing the functions of image and UI display.
  • An input device 280 can be connected to bus 230 via an input/output (I/O) controller 290.
  • I/O controller 290 may utilize one or more of USB, IEEE 1394, or other standards. Multiple input devices may be connected, such as keyboards, mice, and trackpads. Image and video capture devices may also be connected to the system through I/O controller 290 or additional I/O controllers implementing other I/O standards. Networking functionality may be provided by I/O controller 290 or a separate I/O controller.
  • FIG. 3 is a flowchart of an example of a method for distinguishing human errors from errors associated with other causes according to various embodiments.
  • Error information based on user input can be obtained (301) from multiple sources.
  • the error information can be, for example, misspelled words, incorrect use of grammar, etc.
  • the multiple sources can include, for example, devices with different types of HMIs (e.g., mechanical keyboard, virtual keyboard, etc.), different types of computer programs (e.g., texting programs, word processing programs, email programs, etc.), and so forth.
  • An input error can be determined (302) from the error information. For example, a misspelled word can be selected, a violation of a grammar rule can be determined, etc.
  • the method can determine (303) whether the input error is associated with all of the sources or fewer than all of the sources. Errors associated with all of the sources may represent errors caused by human error (such as not knowing the spelling of a word), and errors associated with fewer than all of the sources may represent errors that are caused by issues specific to the associated sources.
  • the associations can be stored (304) in a computer storage memory, such as storage device 103. In this way, a useful data structure (i.e., the stored associations) can be generated.
  • FIG. 4 is a flowchart of an example of a method for determining words to include in a spelling quiz based on distinguishing human errors from HMI-related errors according to various embodiments.
  • Information of user typing on devices with a mechanical keyboard and devices with a virtual keyboard can be received (401).
  • the devices can be, for example, smartphones, tablet computers, personal computers, etc., such as devices 107 A- J described above with respect to FIG. 1.
  • the typing can be, for example, smartphones, tablet computers, personal computers, etc., such as devices 107 A- J described above with respect to FIG. 1.
  • the typing information can be received by a device such as storage server 101 and stored on a storage device, such as storage device 103, for example.
  • the typing information can include, for example, a device identification (ID), which can indicate whether a mechanical or virtual keyboard was used.
  • ID can indicate whether a mechanical or virtual keyboard was used.
  • the typing information received for each device ID can include a history of spelling errors made on the device.
  • spell-checker programs and/or grammar programs running on each device can collect information on the spelling and grammar errors made by the user, which may include, e.g., the misspelled word, how the word was misspelled, the context of the misspelled word, etc.
  • a word that the user has misspelled can be selected (402).
  • the typing information can explicitly identify misspelled words, and selecting a misspelled word can simply select one of the identified words.
  • the typing information may not explicitly identify misspelled words (for example, the typing information may be a history of keystrokes typed by the user), and selecting a word the user has misspelled can include analyzing the history of keystrokes to determine misspelled words.
  • the method determines (403) whether the selected misspelled word was typed and/or written on both types of devices, i.e., the devices with mechanical keyboards and the devices with virtual keyboards, and/or using a handwritten input method such as using a stylus without a virtual keyboard.
  • the method determines (404) not to include the selected word in the quiz, another word the user has misspelled is selected (405), and the process continues.
  • the user may have typed the word "penultimate” only on a mechanical keyboard device and never on a virtual keyboard device.
  • the method can disregard the misspelling of "penultimate” for the purpose of quizzing the user because the method may not be able to determine whether the misspelling is associated with the particular HMI or is associated with human error.
  • various other embodiments may include "penultimate” in the quizzing anyway, and may not even include a determination such a 403.
  • the method determines (406) whether the selected word was misspelled on both types of devices.
  • the method may determine a word was misspelled on a particular type of device if the user misspelled the word one or more times on that type of device.
  • the determination of whether the selected word was misspelled on both types of devices may be based on other criteria. For example, the method may require more than one misspelling (e.g., 5 misspellings) of the word in order for the word to be considered to have been misspelled on the type of device.
  • the determination may include a determination of how frequently the word was misspelled on the type of device (e.g., by dividing the number of times the user misspelled the word with the total number of times the user typed the word), and determining the word was misspelled on the type of device can include determining the frequency of misspelling on the type of device is greater than a threshold frequency.
  • a threshold frequency e.g., a threshold frequency
  • the method adds (407) the selected word to a list of words to include in the spelling quiz, another word the user has misspelled is selected (405), and the process continues.
  • the method can determine that the misspelling of the selected word is associated with human error (i.e., the user has trouble spelling the word) because the error was made regardless of the type of HMI used.
  • the method determines (404) not to include the selected word in the quiz, another word the user has misspelled is selected (405), and the process continues. In this way, for example, the method can determine that the misspelling of the selected word is associated with a particular type of HMI because the error was made only when using that HMI.
  • the quiz word list generated by the method can represent the associations determined above, i.e., association of misspelled words with human error and error associated with a different cause.
  • the quiz word list can be stored, e.g., in storage server 101, and can be used by a spelling quiz application to generate quizzes for the user.
  • FIG. 5 is a flowchart of an example of a method for modifying an auto-suggest function of a grammar checking application based on distinguishing human errors from errors associated with a type of computer program according to various embodiments.
  • Information of user typing in texting programs and typing in word processing programs can be received (501).
  • the grammar information can be received by a device such as storage server 101 and stored on a storage device, such as storage device 103, for example.
  • the grammar information can include, for example, a computer program identification (ID), which can indicate whether a texting program or a word processing program was used.
  • the grammar information received for each type of computer program can include a history of grammar errors made when using the computer program.
  • grammar-checker functions of each computer program can collect information on the grammar errors made by the user, which may include, e.g., the violations of rules of grammar, how the rule was violated, the context of the violation, etc.
  • a grammar rule that the user has violated can be selected (502).
  • the grammar information can explicitly identify violations of grammar rules, and selecting a rule can simply select one of the identified rules that the user violated.
  • the grammar information may not explicitly identify violations of grammar rules (for example, the grammar information may be a history of keystrokes typed by the user), and selecting a rule the user has violated can include analyzing the history of keystrokes to determine grammar violations.
  • the method determines (503) whether the selected grammar rule was violated in both types of computer programs, i.e., texting programs and word processing programs.
  • both types of computer programs i.e., texting programs and word processing programs.
  • analyses may be used to determine whether a grammar rule violation should be considered to have occurred. Analyses may be directed to, for example, identifying a statistically meaningful difference between the errors made in one type of computer program versus the other type of computer program.
  • the method tags (504) the selected grammar rule for autosuggestion in both types of computer programs. Then, another grammar rule the user has violated is selected (505), and the process continues. In this way, for example, the method can determine that the grammar violation is associated with human error (i.e., the user has trouble remembering the rule) because the error was made regardless of the type of computer program used.
  • the method tags (506) the selected grammar rule for autosuggestion only in the type of computer program in which the violation has not occurred. Then, another grammar rule the user has violated is selected (505), and the process continues. For example, the user may violate a particular grammar rule when using texting programs, but may not violate the rule when using word processing programs. This may be because most of the user's texts are communications with friends, and the user prefers to use less formal grammar with friends, while most of the user's word processing documents are for work and require more formal grammar.
  • the method can determine that the violation of the rule is associated with a particular type of computer program, i.e., texting programs, rather than a human error such as not knowing the rule, because the error was made only when using texting programs. Therefore, the method can determine that if the user violates the rule while using a word processing program, that particular violation is likely the result of a genuine mistake (i.e., user error) of the user, in which case an auto-suggestion would be helpful.
  • a genuine mistake i.e., user error
  • the tags generated by the method can represent the associations determined above, i.e., association of grammar rule violations with human error or error associated with a different cause.
  • the tags can be stored, e.g., in storage server 101, and can be used by the grammar-checker functions of each computer program to generate grammar auto-suggestions that are more appropriate for the user's use of the computer program. For example, the auto-suggestion function of the texting program would not bother the user with suggestions for correcting violations of the particular grammar rule that the user typically violates when texting to friends.
  • a computing system such as a general purpose computer through computer-executable instructions (e.g., software, firmware, etc.) stored on a computer-readable medium (e.g., storage disk, memory, etc.) and executed by a computer processor.
  • a computer-readable medium e.g., storage disk, memory, etc.
  • software implementing one or more methods shown in the flowcharts could be stored in storage device 240 and executed by processor 210.
  • various elements shown in the figures may be implemented in various forms of hardware, software or combinations thereof. That is, various elements may be implemented in a combination of hardware and software on one or more appropriately programmed general-purpose devices, which may include a processor, memory and input/output interfaces.
  • the functions of the various elements shown in the figures may be provided through the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software.
  • the functions may be provided by a single dedicated processor, by a single shared processor, or by a plurality of individual processors, some of which may be
  • processor or “controller” should not be construed to refer exclusively to hardware capable of executing software, and may implicitly include, without limitation, digital signal processor (“DSP”) hardware, read only memory (“ROM”) for storing software, random access memory (“RAM”), and nonvolatile storage.
  • DSP digital signal processor
  • ROM read only memory
  • RAM random access memory
  • any switches shown in the figures are conceptual only. Their function may be carried out through the operation of program logic, through dedicated logic, through the interaction of program control and dedicated logic, or even manually, the particular technique being selectable by the implementer as more specifically understood from the context.
  • such phrasing is intended to encompass the selection of the first listed option (A) only, or the selection of the second listed option (B) only, or the selection of the third listed option (C) only, or the selection of the first and the second listed options (A and B) only, or the selection of the first and third listed options (A and C) only, or the selection of the second and third listed options (B and C) only, or the selection of all three options (A and Band C). This may be extended for as many items as listed.
  • any element expressed as a means for performing a specified function is intended to encompass any way of performing that function including, for example, a combination of circuit elements that performs that function, software in any form, including, therefore, firmware, microcode or the like, combined with appropriate circuitry for executing that software to perform the function, etc.
  • the disclosure as defined by such claims resides in the fact that the functionalities provided by the vaiious recited means are combined and brought together in the manner the claims call for. It is thus regarded that any means that can provide those functionalities are equivalent to those shown herein.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Educational Technology (AREA)
  • Educational Administration (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Engineering & Computer Science (AREA)
  • Machine Translation (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

La présente invention vise à distinguer des erreurs associées à une erreur humaine d'erreurs associées à d'autres causes, sur la base de l'analyse d'informations d'erreurs provenant de multiples sources. Des informations d'erreurs basées sur une entrée d'utilisateur sont obtenues à partir d'une pluralité de sources. Une erreur d'entrée est déterminée à partir des informations d'erreur. Il est déterminé si l'erreur d'entrée est associée à toutes les sources ou est associée à moins que toutes les sources. Des informations d'association basées sur une association déterminée sont stockées.
PCT/US2015/068296 2015-12-31 2015-12-31 Identification d'erreurs dans des données d'entrée provenant de sources multiples WO2017116471A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/US2015/068296 WO2017116471A1 (fr) 2015-12-31 2015-12-31 Identification d'erreurs dans des données d'entrée provenant de sources multiples

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/US2015/068296 WO2017116471A1 (fr) 2015-12-31 2015-12-31 Identification d'erreurs dans des données d'entrée provenant de sources multiples

Publications (1)

Publication Number Publication Date
WO2017116471A1 true WO2017116471A1 (fr) 2017-07-06

Family

ID=55221549

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2015/068296 WO2017116471A1 (fr) 2015-12-31 2015-12-31 Identification d'erreurs dans des données d'entrée provenant de sources multiples

Country Status (1)

Country Link
WO (1) WO2017116471A1 (fr)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6889361B1 (en) * 2000-06-13 2005-05-03 International Business Machines Corporation Educational spell checker
US20080052272A1 (en) * 2006-08-28 2008-02-28 International Business Machines Corporation Method, System and Computer Program Product for Profile-Based Document Checking
WO2008053466A2 (fr) * 2006-10-30 2008-05-08 Cellesense Technologies Ltd. Correction d'erreur sensible au contexte de messages texte courts
US20130211824A1 (en) * 2012-02-14 2013-08-15 Erick Tseng Single Identity Customized User Dictionary
WO2014035773A1 (fr) * 2012-08-31 2014-03-06 Microsoft Corporation Correction automatique sensible au contexte

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6889361B1 (en) * 2000-06-13 2005-05-03 International Business Machines Corporation Educational spell checker
US20080052272A1 (en) * 2006-08-28 2008-02-28 International Business Machines Corporation Method, System and Computer Program Product for Profile-Based Document Checking
WO2008053466A2 (fr) * 2006-10-30 2008-05-08 Cellesense Technologies Ltd. Correction d'erreur sensible au contexte de messages texte courts
US20130211824A1 (en) * 2012-02-14 2013-08-15 Erick Tseng Single Identity Customized User Dictionary
WO2014035773A1 (fr) * 2012-08-31 2014-03-06 Microsoft Corporation Correction automatique sensible au contexte

Similar Documents

Publication Publication Date Title
AU2018226398B2 (en) Automated highlighting of key points in content using machine learning to improve document comprehension. using easy to collect human feedback to improve the algorithm
CN105009064B (zh) 使用语言和空间模型的触摸键盘
US10460029B2 (en) Reply information recommendation method and apparatus
CN104718543B (zh) 一种用于建议校正的方法和计算系统
US11194448B2 (en) Apparatus for vision and language-assisted smartphone task automation and method thereof
US20180173698A1 (en) Knowledge Base for Analysis of Text
US20210011595A1 (en) Terminal and method for determining type of input method editor
US10691887B2 (en) Techniques for automatic proofing of textual data
AU2014212844B2 (en) Character and word level language models for out-of-vocabulary text input
US20190205743A1 (en) System and method for detangling of interleaved conversations in communication platforms
US20150309984A1 (en) Learning language models from scratch based on crowd-sourced user text input
CN106202059A (zh) 机器翻译方法以及机器翻译装置
EP2909741A2 (fr) Reconnaissance incrémentale de plusieurs mots
WO2014047161A2 (fr) Sélection adaptative à la posture
CN106663093A (zh) 用于为机器翻译自动地交换语言和/或内容的技术
US20210056113A1 (en) Utilizing unsumbitted user input data for improved task performance
EP2909702B1 (fr) Séparateurs automatiques avec spécificité contextuelle
US10073828B2 (en) Updating language databases using crowd-sourced input
CN105074643A (zh) 非词典字符串的手势键盘输入
US20130346904A1 (en) Targeted key press zones on an interactive display
CN109492085B (zh) 基于数据处理的答案确定方法、装置、终端及存储介质
KR101476229B1 (ko) 수식 입력 방법 및 장치
WO2016191913A1 (fr) Systèmes et procédés pour fournir un agrégateur de nouvelles centré sur les commentaires
CN107797676B (zh) 一种单字输入方法及装置
WO2016155643A1 (fr) Procédé et dispositif d'affichage de mot candidat sur la base d'une entrée

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15828472

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15828472

Country of ref document: EP

Kind code of ref document: A1