CUSTOM LANGUAGE MODELS

• US 20110296374A1
• Filed: 11/05/2008
• Published: 12/01/2011
• Est. Priority Date: 11/05/2008
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Wang, Yong-Gang, Tang, Xiliu, Liu, Yongyan, Ou, Henry, Wu, Jun
• Granted Patent: US 8,826,226 B2
• Time in Patent Office: Days
• US Class Current: 717/104
• CPC Class Codes: G06F 40/216 using statistical methods

Abstract

Systems, methods, and apparatuses including computer program products for generating a custom language model. In one implementation, a method is provided. The method includes receiving a collection of documents; clustering the documents into one or more clusters; generating a cluster vector for each cluster of the one or more clusters; generating a target vector associcated with a target profile; comparing the target vector with each of the cluster vectors; selecting one or more of the one or more clusters based on the comparison; and generating a language model using documents from the one or more selected clusters.

Claims

1. A method comprising:

receiving a collection of documents;

clustering the documents into one or more clusters;

generating a cluster vector for each cluster of the one or more clusters;

generating a target vector associated with a target profile;

comparing the target vector with each of the cluster vectors;

selecting one or more of the one or more clusters based on the comparison; and

generating a language model using documents from the one or more selected clusters.

2. The method of claim 1, where generating a language model includes:

identifying a collection of n-grams from the documents from the one or more selected clusters, each n-gram having a corresponding relative frequency of occurring in the documents from the one or more selected clusters.

3. The method of claim 1, where each cluster is associated with a topic or domain.

4. The method of claim 1, where each cluster vector represents a centroid for a corresponding cluster.

5. The method of claim 4, where clustering the documents includes:

randomly assigning each document to a cluster; and

iteratively calculating a centroid for each cluster and reassigning each document to a cluster based on a nearest centroid, until convergence.

6. The method of claim 1, where the target vector comprises a term frequency vector.

7. The method of claim 1, where the target vector is generated from at least one of a user dictionary or user input.

8. The method of claim 1, where the comparing includes:

calculating a cosine similarity between the target vector and each of the cluster vectors.

9. The method of claim 1, further comprising:

receiving one or more additional documents; and

updating the language model based on the one or more additional documents.

10. A method comprising:

receiving a collection of documents;

clustering the documents into one or more generic clusters;

generating a cluster vector for each cluster of the one or more generic clusters;

generating a target vector associated with a target profile;

comparing the target vector with each of the cluster vectors; and

selecting one or more of the one or more generic clusters based on the comparison.

11. The method of claim 10, further comprising:

generating a language model using documents from the one or more selected generic clusters.

12. A method comprising:

receiving a user input identifying the user;

identifying a user profile corresponding to the user;

using the identified profile to generate a user specific language model; and

sending the user specific language model to a first client.

13. The method of claim 12, further comprising:

sending the user specific language model to a second client.

14. The method of claim 12, further comprising:

receiving a collection of documents;

clustering the documents into one or more clusters;

generating a cluster vector for each cluster of the one or more clusters;

generating a target vector associated with the user profile;

comparing the target vector with each of the cluster vectors; and

selecting one or more of the one or more clusters based on the comparison.

15. The method of claim 14, where generating a user specific language model includes:

identifying a collection of n-grams from the documents from the one or more selected clusters, each n-gram having a corresponding relative frequency of occurring in the documents from the one or more selected clusters.

16. A method comprising:

receiving a first collection of one or more documents;

generating a profile based on the first collection of one or more documents;

receiving a second collection of one or more documents;

generating a custom language model based on the second collection of one or more documents and the profile; and

sending the custom language model to a client.

17. The method of claim 16, where the first collection of one or more documents comprise at least one of a user dictionary or user input representing the user'"'"'s input behavior.

18. A computer program product, encoded on a tangible program carrier, operable to cause data processing apparatus to perform operations comprising:

receiving a collection of documents;

clustering the documents into one or more clusters;

generating a cluster vector for each cluster of the one or more clusters;

generating a target vector associated with a target profile;

comparing the target vector with each of the cluster vectors;

selecting one or more of the one or more clusters based on the comparison; and

generating a language model using documents from the one or more selected clusters.

19. A system, comprising:

a machine-readable storage device including a program product; and

one or more computers operable to execute the program product and perform operations comprising;

receiving a collection of documents;

clustering the documents into one or more clusters;

generating a cluster vector for each cluster of the one or more clusters;

generating a target vector associated with a target profile;

comparing the target vector with each of the cluster vectors;

selecting one or more of the one or more clusters based on the comparison; and

generating a language model using documents from the one or more selected clusters.