Adaptive input interface

• US 8,712,931 B1
• Filed: 06/29/2011
• Issued: 04/29/2014
• Est. Priority Date: 06/29/2011
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Wahlen, Christopher J.
• Primary Examiner(s): Chaki, Kakali
• Assistant Examiner(s): Wu, Fuming
• Application Number:
  US13/172,656
• Time in Patent Office: 1,035 Days
• Field of Search: 706/52
• US Class Current: 706/12
• CPC Class Codes: G06F 3/0237 using prediction or retrieval techniques
• CPC Class Codes: G06F 40/274 Converting codes to words; Guess-ahead of partial word inputs

Abstract

User input on a computing device can be intelligently predicted using one or more learning algorithms. When user text input is received, an input repository comprising a plurality of text items is accessed. The plurality of text items includes a plurality of user-specific text items, which are classified according to a probabilistic model. One or more input predictions are inferred by applying the probabilistic model to the user text input. A suggested text input is presented using an output element of the computing device. The suggested text input is based on the user text input and one or more text items from the input repository. Additional input is received which overrides the suggested text. Any text associated with the user override is then classified within the probabilistic model, thereby updating the input repository based on the user override.

Claims

1. A system for predicting user text input in an electronic environment, comprising:

one or more device processors; and

a memory device including instructions that, when executed by the one or more device processors, cause the system to;

receive, over a communications network, the user text input;

access a text repository storing a plurality of text items classified according to a Bayesian network;

infer one or more input predictions through, at least in part, locating the user text input in the Bayesian network;

transmit data, over the communications network, including a suggested text based at least in part on the user text input and the inferred one or more input predictions;

receive, over the communications network, an indication of a user override of the suggested text, the user override including one or more entries of a backspace key and replacement user text input; and

update a probability of at least one node of the Bayesian network corresponding to the user override,

wherein the probability of the at least one node is based at least in part upon an average probability of one or more other nodes of the Bayesian network when the at least one node is a new node of the Bayesian network, the one or more other nodes corresponding to a same intermediary node of the Bayesian network as the at least one node.

2. The system of claim 1, wherein accessing the text repository comprises retrieving user-specific text items based on a user identifier.

3. The system of claim 1, wherein the user text input comprises one or more alphanumeric characters.

4. A computer-implemented method for intelligently predicting user input entered on a computing device, comprising:

receiving, on the computing device, user text input;

accessing a text repository including a plurality of text items classified according to a probabilistic model;

inferring, using a processor of the computing device, one or more input predictions by applying the probabilistic model to the user text input;

presenting, using a visual output element of the computing device, a suggested text based on the one or more input predictions;

receiving, on the computing device, an indication of a user override of the suggested text; and

updating at least one probability of the probabilistic model, the at least one probability corresponding to the user override,

wherein the at least one probability is based at least in part upon an average probability of one or more other probabilities of the probabilistic model when the user override is a new probability of the probabilistic model, the one or more other probabilities corresponding to the text items that have at least some same text as the user override.

5. The computer-implemented method of claim 4, wherein receiving the indication of the user override of the suggested text comprises detecting an override trigger in the user override, wherein the override trigger comprises an explicit override or an implicit override.

6. The computer-implemented method of claim 4, further comprising:

presenting, using the visual output element of the computing device, a second suggested text related to the user text input and the user override.

7. The computer-implemented method of claim 4, wherein the user text input comprises at least one non-space character followed by a space or punctuation character.

8. The computer-implemented method of claim 4, wherein accessing the text repository comprises retrieving user-specific text items from a data store on the computing device.

9. The computer-implemented method of claim 4, wherein the text repository stores data representing a user-specific collection of vocabulary terms, and wherein the suggested text comprises one or more vocabulary terms or a correction regarding spelling.

10. The computer-implemented method of claim 4, wherein the text repository comprises a set of user-specific grammar rules, and wherein the suggested text comprises a correction regarding capitalization, punctuation, or grammar.

11. The computer-implemented method of claim 4, wherein the probabilistic model comprises a Bayesian network, an n-gram model, or a hidden Markov model.

12. The computer-implemented method of claim 4, wherein updating the text repository based on the user override comprises storing the user override in the text repository.

13. The computer-implemented method of claim 4, wherein applying the probabilistic model to the user text input comprises determining the one or more input predictions according to a posterior probability distribution.

14. The computer-implemented method of claim 13, wherein classifying the user override within the probabilistic model comprises updating a prior probability distribution of the probabilistic model.

15. The computer-implemented method of claim 13, wherein classifying the user override within the probabilistic model comprises updating a smoothing algorithm associated with the plurality of text items in the text repository.

16. The computer-implemented method of claim 4, wherein presenting the suggested text comprises modifying the user text input to an input prediction with at least a minimum probability.

17. A computing device, comprising:

a processor;

at least one input detection element;

at least one display element; and

a memory device including instructions operable to be executed by the processor to perform a set of actions, enabling the computing device to;

receive user text input via the at least one input detection element;

compare the user text input against a text repository including a plurality of text items classified according to a probabilistic model to determine at least one predicted input string;

provide the at east one predicted input string for display on the at least one display element;

receive, via the at least one input detection element, an indication of a user override of the at least one predicted input string; and

update at least one probability of the probabilistic model, the at least one probability corresponding to the user override,

wherein the at least one probability is based at least in part upon an average probability of one or more other probabilities of the probabilistic model when the user override is a new probability of the probabilistic model, the one or more other probabilities corresponding to the text items that have at least some same text as the user override.

18. The computing device of claim 17, wherein the user text input is detected through at least one of an applied pressure, physical contact, gesture, or audio command of a user with respect to the computing device.

19. The computing device of claim 17, wherein the memory device further includes instructions that, when executed, cause the computing device to:

present, using the at least one display element of the computing device, a second suggested text related to the user text input and the user override.

20. The computing device of claim 17, wherein the at least one display element of the computing device includes a display screen or a video output, and wherein presenting a suggested text related to the user text input comprises displaying the suggested text on the display screen or using the video output of the computing device.

21. The computing device of claim 17, wherein the instructions to provide the at least one predicted input string further enable the computing device to:

display a pop-up text area including one or more suggested terms that are able to be selected.

22. One or more non-transitory computer-readable storage media having collectively stored thereon instructions that, when executed by one or more processors of a network system, cause the network system to at least:

obtain user text input entered by a user into an electronic device;

determine at least one input prediction by applying a probabilistic model to the user text input and selecting one or more associated text items from a text repository classified according to the probabilistic model;

transmit information including the at least one input suggestion for presentation to the user on the electronic device;

receive an indication of a user override of the at least one input suggestion on the electronic device; and

update at least one probability of the probabilistic model, the at least one probability corresponding to the user override,

wherein the at least one probability is based at least in part upon an average probability of one or more other probabilities of the probabilistic model when the user override is a new probability of the probabilistic model, the one or more other probabilities corresponding to the one or more associated text items that have at least some same text as the user override.

23. The one or more non-transitory computer-readable storage media of claim 22, wherein the instructions when executed further cause the network system to:

receive a subsequent instance of the user text input entered by the user; and

determine at least one input prediction utilizing any text associated with the user override.

24. The one or more non-transitory computer-readable storage media of claim 22, wherein the instructions when executed further cause the network system to:

receive a subsequent instance of the user text input entered by the user; and

determine at least one different input prediction based upon information stored for the user override.

25. The one or more non-transitory computer-readable storage media of claim 22, wherein an output element of the electronic device to present the at least one input suggestion comprises a speaker or an audio output, and wherein the instructions when executed further cause the network system to:

transmit sound representing a voice enunciating the at least one input suggestion through the speaker or the audio output of the electronic device.