Unlearning techniques for adaptive language models in text entry

Patent metadata

Field	Value
Publication number	US-9703394-B2
Application number	US-201514873147-A
Country	US
Kind code	B2
Filing date	Oct 1, 2015
Priority date	Mar 24, 2015
Publication date	Jul 11, 2017
Grant date	Jul 11, 2017

Abstract

In some examples, a method includes outputting a graphical keyboard ( 120 ) for display and responsive to receiving an indication of a first input ( 124 ), determining a new character string that is not included in a language model. The method may include adding the new character string to the language model and associating a likelihood value with the new character string. The method may include, responsive to receiving an indication of a second input, predicting the new character string, and responsive to receiving an indication of a third input that rejects the new character string, decreasing the likelihood value associated with the new character string.

First claim

What is claimed is:

1. A method comprising:
outputting, by a computing device and for display at an output component operatively coupled to the computing device, a graphical user interface that includes a text edit region and a graphical keyboard;
receiving, by the computing device, an indication of a first input detected at a first location of a presence-sensitive input component operatively coupled to the computing device, wherein the first location of the presence-sensitive input component corresponds to a region of the output component at which the graphical keyboard is displayed;
determining, by the computing device, based at least in part on the indication of the first input, a new character string that is not included in a language model of the computing device;
adding, by the computing device, to the language model, the new character string and a corresponding likelihood value associated with the new character string;
receiving, by the computing device, an indication of a second input detected at a second location of the presence-sensitive input component, wherein the second location of the presence-sensitive input component corresponds to the region of the output component at which the graphical keyboard is displayed;
predicting, based on the indication of the second input and the likelihood value associated with the new character string, the new character string and an input context associated with the second input;
responsive to predicting the new character string, outputting, by the computing device, for display within the text edit region, the new character string;
receiving, by the computing device, an indication of a third input detected at a third location of the presence-sensitive input component, wherein the third location of the presence-sensitive input component corresponds to the region of the output component at which the graphical keyboard is displayed; and
responsive to determining that the third input deletes the new character string from the text input region after being output for display:
removing, by the computing device, from the text input region, the new character string; and
decreasing, by the computing device, at a rate or magnitude that is adapted to the input context, the likelihood value associated with the new character string without removing the new character string from the language model.

2. The method of claim 1 , wherein associating the likelihood value with the new character string comprises storing the new character string in at least one n-gram of the language model with the likelihood value, such that the likelihood value corresponds to the at least one n-gram in the language model, the method further comprising:
responsive to determining that the third input deletes the new character string:
determining a plurality of n-grams in the language model that include the new character string; and
decreasing a respective likelihood value that corresponds to each respective n-gram of the plurality of n-grams.

3. The method of claim 2 , further comprising:
responsive to determining that a first likelihood value that corresponds a higher-order n-gram of the language model is greater than a second likelihood value that corresponds to a lower-order n-gram, modifying the language model to at least:
remove the higher-order n-gram, or
update the first likelihood value to be less than or equal to the second likelihood value,

wherein the higher-order n-gram includes more character strings than the lower-order n-gram.

4. The method of claim 1 , further comprising:
wherein the input context comprises one or more previously inputted character strings that are outputted for display when predicting the new character string,
wherein decreasing the likelihood value associated with the new character string comprises decreasing the likelihood value associated with the new character string for only the input context.

5. The method of claim 1 , wherein a rate or magnitude of increasing the likelihood value that corresponds to the new character string in response to learning the new character string is different than the rate or magnitude that is adapted to the input context.

6. The method of claim 1 , further comprising:
after decreasing the likelihood value associated with the new character string, receiving, by the computing device, an indication of a fourth input detected at a fourth location of the presence-sensitive input component, wherein the fourth location of the presence-sensitive input component corresponds to the region of the output component at which the graphical keyboard is displayed;
responsive to predicting, based on the indication of the fourth input and the likelihood value associated with the new character string, the new character string, outputting, by the computing device, for display within the text edit region, the new character string;
receiving, by the computing device, an indication of a fifth input detected at a fifth location of the presence-sensitive input component, wherein the fifth location of the presence-sensitive input component corresponds to the region of the output component at which the graphical keyboard is displayed; and
responsive to determining that the fifth input deletes the new character string from the text edit region after being output for display:
removing, by the computing device, from the text input region, the new character string; and
decreasing, by the computing device, the likelihood value associated with the new character string without removing the new character string from the language model.

7. The method of claim 1 , further comprising:
after decreasing the likelihood value associated with the new character string and responsive to determining that the likelihood value is a zero or null value, removing, by the computing device, the new character string from the language model.

8. A computing device comprising:
one or more computer processors;
an output component;
a presence-sensitive input component; and
a memory comprising instructions that when executed cause the one or more computer processors to:
output, for display at the output component, a graphical user interface that includes a text edit region and a graphical keyboard;
receive an indication of a first input detected at a first location of the presence-sensitive input component, wherein the first location of the presence-sensitive input component corresponds to a region of the output component at which the graphical keyboard is displayed;
determine, based at least in part on the indication of the first input, a new character string that is not included in a language model of the computing device;
add, to the language model, the new character string and a corresponding likelihood value associated with the new character string;
receive an indication of a second input detected at a second location of the presence-sensitive input component, wherein the second location of the presence-sensitive input component corresponds to the region of the output component at which the graphical keyboard is displayed;
predict, based on the indication of the second input and the likelihood value associated with the new character string, the new character string and an input context associated with the second input;
responsive to predicting the new character string, output, for display within the text edit region, the new character string;
receive an indication of a third input detected at a third location of the presence-sensitive input component, wherein the third location of the presence-sensitive input component corresponds to the region of the output component at which the graphical keyboard is displayed; and
responsive to determining that the third input deletes the new character

Assignees

• Google Inc

Inventors

• Ouyang Yu
• Zhai Shumin

Classifications

• G06F2203/04808

  Several contacts: gestures triggering a specific function, e.g. scrolling, zooming, right-click, when the user establishes several contacts with the surface simultaneously; e.g. using several fingers or a combination of fingers and pen · CPC title

• G06F3/0237 Primary

  using prediction or retrieval techniques · CPC title

• G06F40/00

  Handling natural language data (speech analysis or synthesis, speech recognition G10L) · CPC title

• G06F40/166

  Editing, e.g. inserting or deleting · CPC title

• G06F40/216

  using statistical methods · CPC title