System for improving autocompletion of text input

1. A method comprising:

after receiving, an indication of user input that selects one or more textual characters, receiving, by a computing device, an indication of user input that selects an end-of-word identifier;

determining, by the computing device, based at least in part on the one or more textual characters and in response to receiving the end-of-word identifier, a first auto-complete word suggestion from a plurality of auto-complete word suggestions, wherein the first auto-complete word suggestion is determined to be more likely to be a correct word suggestion than a second auto-complete word suggestion from the plurality of auto-complete word suggestions;

outputting, by the computing device, for display, the first auto-complete word suggestion that replaces the one or more textual characters followed by the end-of-word identifier;

receiving, by the computing device, an indication of user input that deletes the end-of-word identifier, wherein the indication of user input that deletes the end-of-word identifier comprises a combination of a selection of both a backspace key and a spacebar key; and

responsive to receiving the indication of user input that deletes the end-of-word identifier, outputting, by the computing device, for display, the second auto-complete word suggestion, wherein the second auto-complete word suggestion replaces the first auto-complete word suggestion;

wherein the end-of-word identifier is a first end-of-word identifier and outputting the second auto-complete word suggestion comprises outputting, by the computing device, for display, the second auto-complete word suggestion followed by a second end-of-word identifier replacing the first auto-complete word suggestion and the first end-of-word identifier.

2. The method of claim 1 , further comprising:

determining, by the computing device, a probability that the second auto-complete word suggestion represents the one or more textual characters; and

determining, by the computing device, that the probability satisfies a threshold, wherein outputting the second auto-complete word suggestion is responsive to determining that the probability satisfies the threshold.

3. The method of claim 1 , further comprising:

storing, by the computing device, at least some state information to identify the plurality of auto-complete word suggestions; and

responsive to receiving the indication of user input that deletes the end-of-word identifier, retrieving, by the computing device, the second auto-complete word suggestion from the stored at least some state information to replace the first auto-complete word suggestion.

4. The method of claim 1 , further comprising:

prior to outputting the first auto-complete word suggestion and the end-of-word identifier for display, outputting, by the computing device, for display, the one or more textual characters, wherein outputting the first auto-complete word suggestion of the plurality of auto-complete word suggestions comprises replacing, by the computing device, the one or more textual characters with the first auto-complete word suggestion and the end-of-word identifier.

5. The method of claim 1 , further comprising:

determining, by the computing device, a respective probability that each auto-complete word suggestion of the plurality of auto-complete word suggestions represents the one or more textual characters,

wherein the respective probability of the first auto-complete word suggestion exceeds the respective probability of the second auto-complete word suggestion.

6. The method of claim 5 , further comprising:

determining, by the computing device, using an n-gram language model, the respective probability of at least one auto-complete word suggestion of the plurality of auto-complete word suggestions.

7. The method of claim 5 , further comprising:

determining, by the computing device, based at least in part on a similarity coefficient between the at least one auto-complete word suggestion and one or more words in a lexicon, the respective probability of at least one auto-complete word suggestion of the plurality of auto-complete word suggestions.

8. The method of claim 1 , wherein the end-of-word identifier comprises a representation of a whitespace character, a punctuation character, or a null value character.

9. The method of claim 1 , wherein:

the indication of user input that deletes the first end-of-word identifier is a first indication of user input to that deletes the first end-of-word identifier,

the plurality of auto-complete word suggestions comprises a third auto-complete word suggestion that is less likely to be a correct word suggestion based on the one or more textual characters than both the first auto-complete word suggestion and the second auto-complete word suggestion, and

wherein the method further comprises:

receiving, by the computing device, a second indication of user input that deletes the second end-of-word identifier; and

responsive to receiving the second indication of user input, outputting, by the computing device, for display, the third auto-complete word suggestion, wherein the third auto-complete word suggestion replaces the second auto-complete word suggestion.

10. The method of claim 1 , wherein the indication of user input that selects the one or more textual characters is a first indication of user input that selects the one or more textual characters, the method further comprising:

receiving, by the computing device, a second indication of user input that selects the one or more textual characters;

determining, by the computing device, based on the second indication of user input that selects the one or more textual characters, that the one or more textual characters are correct; and

responsive to determining that the one or more textual characters are correct, outputting, by the computing device, for display, the one or more textual characters by replacing the second auto-complete word suggestion.

11. The method of claim 1 , wherein the indication of user input that selects the one or more textual characters, the indication of user input that selects the end-of-word identifier, and the indication of user input that deletes the end-of-word identifier are each received at an input device comprising a presence-sensitive screen, the method further comprising:

outputting, by the computing device, for display at the presence-sensitive screen, a graphical keyboard comprising a plurality of keys, wherein the indication of user input that selects the one or more textual characters, the indication of user input that selects the end-of-word identifier, and the indication of user input that deletes the end-of-word identifier each comprise respective representations of selections of one or more of the plurality of keys.

12. The method of claim 1 , wherein outputting the second auto-complete word suggestion comprises outputting, by the computing device, for display at a location of a presence-sensitive screen that corresponds to a portion of the presence-sensitive screen at which the first auto-complete word suggestion was displayed, the second auto-complete word suggestion.

13. The method of claim 1 , wherein the indication of user input that selects the one or more textual characters, the indication of user input that selects the end-of-word identifier, and the indication of user input that deletes the end-of-word identifier comprise a single indication of a non-tap gesture detected at a location of a presence-sensitive screen.

14. The method of claim 1 , further comprising:

determining, by the computing device, a highest probability auto-complete word suggestion of the plurality of auto-complete word suggestions that is most likely to be the correct word suggestion based on the one or more textual characters, wherein the first auto-complete word suggestion corresponds to the highest probability auto-complete word suggestion of the plurality of auto-complete word suggestions.

15. A non-transitory computer-readable storage medium encoded with instructions that, when executed, cause at least one processor of a computing device to:

after receiving, by the computing device, an indication of user input that selects one or more textual characters, receive an indication of user input that selects an end-of-word identifier;

determine, by the computing device, based at least in part on the one or more textual characters and in response to receiving the end-of-word identifier, a first auto-complete word suggestion from a plurality of auto-complete word suggestions, wherein the first auto-complete word suggestion is determined to be more likely to be a correct word suggestion than a second auto-complete word suggestion from the plurality of auto-complete word suggestions;

output, by the computing device, for display, the first auto-complete word suggestion that replaces the one or more textual characters followed by the end-of-word identifier;

receive, by the computing device, an indication of user input that deletes the end-of-word identifier, wherein the indication of user input that deletes the end-of-word identifier comprises a combination of a selection of both a backspace key and a spacebar key; and

responsive to receiving the indication of user input that deletes the end-of-word identifier, output, by the computing device, for display, the second auto-complete word suggestion, wherein the second auto-complete word suggestion replaces the first auto-complete word suggestion;

wherein the end-of-word identifier is a first end-of-word identifier and outputting the second auto-complete word suggestion comprises outputting, by the computing device, for display, the second auto-complete word suggestion followed by a second end-of-word identifier replacing the first auto-complete word suggestion and the first end-of-word identifier.

16. A computing system comprising:

at least one processor;

an input device and a display device each operatively coupled to the at least one processor; and

at least one module operable by the at least one processor to:

after receiving, at the input device, an indication of user input that selects one or more textual characters, receive, at the input device, an indication of user input that selects an end-of-word identifier;

determine, based at least in part on the one or more textual characters and in response to receiving the end-of-word identifier, a first auto-complete word suggestion from a plurality of auto-complete word suggestions, wherein the first auto-complete word suggestion is determined to be more likely to be a correct word suggestion than a second auto-complete word suggestion from the plurality of auto-complete word suggestions;

output, for display at the display device, the first auto-complete word suggestion that replaces the one or more textual characters followed by the end-of-word identifier;

receive, at the input device, an indication of user input that deletes the end-of-word identifier, wherein the indication of user input that deletes the end-of-word identifier comprises a combination of a selection of both a backspace key and a spacebar key; and

responsive to receiving the indication of user input that deletes the end-of-word identifier, output, for display at the display device, the second auto-complete word suggestion, wherein the second auto-complete word suggestion replaces the first auto-complete word suggestion;

wherein the end-of-word identifier is a first end-of-word identifier and outputting the second auto-complete word suggestion comprises outputting, by the computing device, for display, the second auto-complete word suggestion followed by a second end-of-word identifier replacing the first auto-complete word suggestion and the first end-of-word identifier.

17. The method of claim 1 , wherein the indication of user input that selects the end-of-word identifier comprises a representation of a selection of a spacebar key and the indication of user input that deletes the end-of-word identifier comprises a representation of a selection of a backspace key after the selection of the spacebar key.