SYSTEM AND METHOD FOR INPUTTING TEXT INTO ELECTRONIC DEVICES

• US 20130041857A1
• Filed: 03/04/2011
• Published: 02/14/2013
• Est. Priority Date: 03/04/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Touchtype Limited (Microsoft Corporation)
• Inventors: Medlock, Benjamin, Reynolds, Jonathan
• Granted Patent: US 9,052,748 B2
• Time in Patent Office: Days
• US Class Current: 706/14
• 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

The present invention relates generally to a system and method for the reordering of text predictions. More particularly, the system and method reorders the text predictions based on modified probability values, wherein the probability values are modified according to the likelihood that a given text prediction will occur in the text inputted by a user.

Claims

1. A system for generating text input in a computing device, comprising;

a predictor configured to receive text input into the device by a user and to generate text predictions having probability values;

a mechanism configured to receive the text predictions and the user text input and to generate modified probability values based upon the user text input and the text predictions for presentation to the user.

3. The system according to claim 1, wherein the mechanism comprises a vector map comprising a context vector for terms in a set of documents, wherein the context vector comprises a sum of index vectors for all documents of the set of documents in which that term occurs.

4. The system according to claim 3, wherein the vector map is a Random Indexing Term-Vector Map.

5. The system according to claim 4, wherein the mechanism is configured to generate a set of Prediction Vectors, comprising a context vector retrieved from the vector map for each text prediction that has an equivalent in the vector map.

6. The system according to claim 5, wherein if the text prediction is a phrase, the mechanism is configured to generate a Prediction Vector comprising an average of the context vectors corresponding to each term within the phrase.

7. The system according to claim 5, wherein the mechanism is configured to retrieve, from the vector map, a context vector for each term in the text input by the user that has an equivalent in the vector map and to generate an Average Document Vector by determining the average of the context vectors retrieved.

8. The system according to claim 7, wherein the mechanism further comprises a Similarity Module which is configured to determine similarity between the Average Document Vector and each of the Prediction Vectors, to generate a similarity value associated with each text prediction.

10. The system according to claim 8, wherein the mechanism further comprises a weighting module which is configured to modify the probability values associated with each text prediction on the basis of the similarity values.

12. The system according to claim 3, wherein the mechanism is configured to update the vector map by assigning a new index vector to a completed text sequence input by a user and by adding the new index vector to the sum of index vectors for each term contained in that completed text sequence.

2. (canceled)

9. (canceled)

11. (canceled)

13-17. -17. (canceled)

18. A method of generating text input for a computing device from user input, comprising:

generating text predictions based upon user input;

generating a modified probability value for each of the text predictions based upon the user input and the text predictions; and

presenting the text predictions to the user for selection.

20. The method according to claim 18, wherein generating a modified probability value for each of the text predictions comprises generating a vector map which comprises a context vector for each term in a set of documents, wherein the context vector comprises a sum of index vectors for all documents of the set of documents in which that term occurs.

21. The method according to claim 20, wherein the step of generating a vector map further comprises filtering out from the vector map terms with a very high occurrence rate across the set of documents or stopwords.

22. The method according to claim 20, wherein the vector map is a Random Indexing Term-Vector Map, and generating a vector map comprises generating context vectors using Random Indexing.

23. The method according to claim 20, further comprising generating a set of Prediction Vectors, by retrieving from the vector map a context vector for each text prediction that has an equivalent in the vector map.

24. The method according to claim 23, wherein if the text prediction is a phrase, the generation of a Prediction Vector further comprises averaging the context vectors corresponding to each term within the phrase.

25. The method according to claim 23, further comprising retrieving, from the vector map, a context vector for each term in the user text input which has an equivalent in the vector map and generating an Average Document Vector by calculating the average of the context vectors retrieved for each term in the text input by the user.

26. The method according to claim 25, further comprising calculating the similarity between the Average Document Vector and each of the Prediction Vectors and providing a similarity value associated with each text prediction.

28. The method according to claim 26, further comprising modifying the probabilities associated with each text prediction on the basis of the similarity values.

30. The method according to claim 20, wherein once a user has entered a complete text sequence, the method further comprises adding that text sequence to the set of documents.

31. The method according to claim 30, further comprising updating the vector map by assigning a new index vector to the completed text sequence and by adding the new index vector to the sum of index vectors for each term contained in that completed text sequence.

19. (canceled)

27. (canceled)

29. (canceled)

32-35. -35. (canceled)