USER-CENTRIC SOFT KEYBOARD PREDICTIVE TECHNOLOGIES

• US 20110202876A1
• Filed: 03/22/2010
• Published: 08/18/2011
• Est. Priority Date: 02/12/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Almog, Itai, Gunawardana, Asela J, Sundararajan, Parthasarathy, Rudchenko, Dmytro, Badger, Eric Norman, Linerud, Drew Elliot, Paek, Timothy S.
• Granted Patent: US 8,782,556 B2
• Time in Patent Office: Days
• US Class Current: 715/816
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 3/0237 using prediction or retrieval techniques
• CPC Class Codes: G06F 3/048 Interaction techniques based on graphical user interfaces [GUI]
• CPC Class Codes: G06F 3/04883 for inputting data by handwriting, e.g. gesture or text
• CPC Class Codes: G06F 3/04886 by partitioning the display area of the touch-screen or the surface of the digitising tablet into independently controllable areas, e.g. virtual keyboards or menus
• CPC Class Codes: G06F 40/166 Editing, e.g. inserting or deleting
• CPC Class Codes: G06F 40/247 Thesauruses; Synonyms
• CPC Class Codes: G06F 40/274 Converting codes to words; Guess-ahead of partial word inputs
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

An apparatus and method are disclosed for providing feedback and guidance to touch screen device users to improve text entry user experience and performance by generating input history data including character probabilities, word probabilities, and touch models. According to one embodiment, a method comprises receiving first input data, automatically learning user tendencies based on the first input data to generate input history data, receiving second input data, and generating auto-corrections or suggestion candidates for one or more words of the second input data based on the input history data. The user can then select one of the suggestion candidates to replace a selected word with the selected suggestion candidate.

Claims

1. A method, comprising:

receiving first input data;

automatically learning user tendencies based on the first input data to generate input history data;

receiving second input data; and

based on the input history data and the second input data, generating auto-corrections or suggestion candidates for the second input data.

2. The method of claim 1, wherein the receiving first input data comprises receiving a completion event indicating that the first input data is complete, wherein the completion event is associated with the sending of a text message, instant message, or email.

3. The method of claim 1, wherein the automatically learning user tendencies comprises analyzing user-specific web content associated with a user.

4. The method of claim 1, wherein the input history data includes touch position data for one or more key presses received using a touch screen keyboard.

5. The method of claim 1, wherein:

the first input data comprises a word; and

the input history data comprises data representing the last time a word was used.

6. The method of claim 1, wherein the input history data comprises typing speed data for one or more keystrokes received using a hardware or touch screen keyboard.

7. The method of claim 1, wherein the input history data comprises orientation data representing the orientation of a touch screen while receiving the first input data.

8. The method of claim 1, wherein:

the first input data and the second input data are received at a computing cloud from a text entry device;

the text entry device is at a location remote from the cloud;

the automatically learning user tendencies occurs in the cloud; and

the generated auto-corrections or suggestion candidates are provided from the cloud to the text entry device.

9. The method of claim 1, wherein:

the receiving second input data further comprises resizing one or more target areas associated with one or more associated keys on a touch screen keyboard based on one or more characters previously received in the second input data; and

one or more characters of the second input data are determined at least in part using the resized target areas.

10. The method of claim 9, wherein the resizing the one or more target areas includes using a generic touch model based on previously-collected touch screen keyboard typing data for plural users, the generic touch model including one or more child probability scores for one or more input data sequences for the typing data.

11. A computer-readable storage device storing computer-readable instructions that when executed cause a computer to perform the method of claim 1.

12. A method, comprising:

receiving first input data associated with an input scope of a mobile device, wherein the first input data comprises one or more words;

automatically learning user tendencies using the first input data to generate user-specific input history data;

receiving second input data from a touch screen keyboard;

analyzing the second input data and at least a portion of the user-specific input history data; and

based on the analyzing, providing one or more user-specific auto-corrections or suggestion candidates for the second input data to the touch screen keyboard.

13. The method of claim 12, wherein the user-specific input history data includes data indicating that at least one input word of the first input data was previously auto-corrected and that the auto-correction was undone.

14. The method of claim 12, wherein the user-specific auto-corrections comprise suppressing an auto-correction based on comparing a threshold value to a score generated by combining first data comprising a weighted probability based on one or more candidate sources with second data comprising a weighted edit distance based on one more edit distance penalty categories.

15. The method of claim 12, wherein:

the analyzing comprises combining a weighted probability for at least one of the words based on one or more candidate sources with a weighted edit distance based on one or more edit distance penalty categories; and

at least one of the one or more candidate sources is based on user-specific input history data.

16. The method of claim 12, wherein the receiving second input data further comprises resizing one or more target areas associated with one or more associated keys on the touch screen keyboard, wherein the resizing is based on one or more characters previously received in the second input data.

17. The method of claim 16, wherein the resizing is based at least in part on the user-specific input history data.

18. The method of claim 12, wherein:

the first input data and the second input data are received at a computing cloud from the mobile device;

the mobile device is at a location remote from the cloud;

the automatically learning user tendencies occurs in the cloud;

the analyzing further comprises combining input history data from two or more distinct users; and

the generated auto-corrections or suggestion candidates are provided from the cloud to the mobile device.

19. A computer-readable storage device storing computer-readable instructions that when executed cause a computer to perform the method of claim 12.

20. A computer-readable storage device storing computer-readable instructions that when executed cause a computer to perform a method, the method comprising:

using a touch screen keyboard coupled with a mobile device, receiving first user input data associated with an input scope, the first user input data comprising one or more characters;

using the text entry device, receiving a completion event associated with the input scope;

after the receiving, automatically analyzing the first user input data and storing user-specific input history data based on the analyzing, wherein the stored user-specific input history data includes an indication of the input scope;

using the touch screen keyboard, receiving second user input data;

determining a weighted probability determined from one or more candidate source data selected from a group including at least one of the following;

the user-specific input history data, the second user input data, dictionary probability data, user-specific touch model data, and generic touch model data;

determining a weighted edit distance determined by calculating one or more edit distances selected from a group of penalty categories including at least one of the following;

insertion, wildcard, deletion, transposition, non-fuzzy substitution, fuzzy substitution, case change, and accent;

automatically determining one or more suggestion candidates based on a score determined by combining the weighted probability and the weighted edit distance; and

displaying the suggestion candidates on the touch screen keyboard.