Neural network for keyboard input decoding
In some examples, a computing device includes at least one processor; and at least one module, operable by the at least one processor to: output, for display at an output device, a graphical keyboard; receive an indication of a gesture detected at a location of a presence-sensitive input device, wherein the location of the presence-sensitive input device corresponds to a location of the output device that outputs the graphical keyboard; determine, based on at least one spatial feature of the gesture that is processed by the computing device using a neural network, at least one character string, wherein the at least one spatial feature indicates at least one physical property of the gesture; and output, for display at the output device, based at least in part on the processing of the at least one spatial feature of the gesture using the neural network, the at least one character string.
1. A computing device comprising at least one processor configured to:
output for display a graphical keyboard that includes a plurality of keys;
receive an indication of at least one user input at a location of a presence-sensitive input device, the location of the presence-sensitive input device corresponding to a location of the graphical keyboard, the at least one user input including one or more current features; and
output, for display in a suggested character string region, a predicted word or word phrase, the predicted word or word phrase determined using a model that is trained to predict a word or word phrase based on the one or more current features, the model comprising a neural network.
2. The computing device of claim 1 , wherein the neural network is a recurrent neural network, the recurrent neural network comprising a Long Short Term Memory network.
3. The computing device of claim 1 , wherein the at least one user input comprises one or more tap gestures, one or more continuous gestures, or a combination of one or more tap gestures and one or more continuous gestures.
4. The computing device of claim 1 , wherein the predicted word or word phrase comprises at least one of a word prediction, an auto-correction, or a suggestion.
5. The computing device of claim 1 , wherein the one or more current features include at least one of a spatial feature, a temporal feature, a lexical feature, or a contextual feature of the at least one user input.
6. The computing device of claim 5 , wherein the spatial feature indicates at least one of a location of the at least one user input, a speed of the at least one user input, a direction of the at least one user input, a curvature of the at least one user input, any keys of the plurality of keys traversed by the at least one user input, or a type of the at least one user input.
7. The computing device of claim 5 , wherein the temporal feature indicates at least one of an epoch time of the at least one user input or an international standard notation time of the at least one user input.
8. The computing device of claim 5 , wherein the contextual feature indicates at least one of a setting or state of information maintained by the computing device, an identity of a user of the computing device, a geolocation of the computing device, an environment or climate of the computing device, audible or visual information obtained by the computing device, sensor information obtained by the computing device, a type of input field for displaying the predicted word or word phrase, or an application for receiving the at least one user input.
9. The computing device of claim 1 , wherein the at least one processor is further configured to:
determine additional features from another user input to the graphical keyboard;
update the model to input the additional features to the model that already received the one or more current features; and
output, for display in the suggested character string region, another predicted word or phrase in response to providing the additional features to the model.
10. The computing device of claim 1 , wherein the computing device comprises a mobile phone, a tablet computer, a laptop computer, a desktop computer, a gaming device, a media player, an e-book reader, a smart watch, or a television platform.
11. A non-transitory computer-readable storage medium comprising instructions that, when executed, configure one or more processors of a computing device to:
output for display a graphical keyboard that includes a plurality of keys;
receive an indication of at least one user input at a location of a presence-sensitive input device, the location of the presence-sensitive input device corresponding to a location of the graphical keyboard, the at least one user input including one or more current features; and
output, for display in a suggested character string region, a predicted word or word phrase, the predicted word or word phrase determined using a model that is trained to predict a word or word phrase based on the one or more current features, the model comprising a neural network.
12. The non-transitory computer-readable storage medium of claim 11 , wherein the neural network is a recurrent neural network, the recurrent neural network comprising a Long Short Term Memory network.
13. The non-transitory computer-readable storage medium of claim 11 , wherein, the neural network is trained based on features corresponding to user input, to the user inputs indicative of different selections of the plurality of keys for entering word phrases and sentences.
14. The non-transitory computer-readable storage medium of claim 13 , wherein the neural network is further trained to provide probabilities for one or more letters or words based on a context associated with the computing device.
15. The non-transitory computer-readable storage medium of claim 13 , wherein the neural network is further trained based on an identity of a user of the computing device, the identity indicative of a frequency that the predicted word or word phrase occurs in a vocabulary associated with an identity of a user of the computing device.
16. The non-transitory computer-readable storage medium of claim 13 , wherein the neural network is further trained based on a frequency that the predicted word or word phrase occurs in a vocabulary associated with an identity of an intended recipient of text including the predicted word or word phrase.
17. The non-transitory computer-readable storage medium of claim 13 , wherein the neural network is further trained based on a frequency that the predicted word or word phrase occurs in a language context including the word phrases and sentences.
18. The non-transitory computer-readable storage medium of claim 13 , wherein the neural network is configured to execute at one or more processors associated with the computing device.
19. The non-transitory computer-readable storage medium of claim 11 , wherein the at least one user input comprises one or more tap gestures, one or more continuous gestures, or a combination of one or more tap gestures and one or more continuous gestures.
20. The non-transitory computer-readable storage medium of claim 11 , wherein the predicted word or word phrase comprises at least one of a word prediction, an auto-correction, or a suggestion.