Cross-channel clusters of information

• US 8,543,577 B1
• Filed: 03/02/2011
• Issued: 09/24/2013
• Est. Priority Date: 03/02/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Ben-Artzi, Aner, Buryak, Kirill, Lewis, Glenn M., Peng, Jun, Benbarak, Nadav
• Primary Examiner(s): LE, UYEN T
• Application Number:
  US13/038,835
• Time in Patent Office: 937 Days
• Field of Search: None
• US Class Current: 707/738
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification

Abstract

A computer-implemented method includes receiving, by one or more computer systems, first information from a first channel and second information from a second channel; merging the first information with the second information; applying an unsupervised clustering model to the merged information; and generating, based on results of the applying, a cross-channel cluster, the cross-channel cluster including (i) a portion of the first information associated with a subject matter, and (ii) a portion of the second information associated with the subject matter.

Claims

1. A computer-implemented method comprising:

receiving, by one or more computer systems, first information from a first type of channel and second information from a second type of channel, with the first type of channel differing from the second type of channel;

merging the first information received from the first type of channel with the second information received from the second type of channel that differs from the first type of channel;

applying an unsupervised clustering model to the merged information; and

generating, based on results of the applying, a cross-channel cluster, the cross-channel cluster comprising (i) a portion of the first information received from the first type of channel associated with a subject matter, and (ii) a portion of the second information received from the second type of channel that differs from the first type of channel associated with the subject matter, wherein the cross-channel cluster comprises a cluster of information that is received from at least two different types of channels.

2. The computer-implemented method of claim 1, further comprising:

applying a supervised labeling model to identify a label for the cross-channel cluster.

3. The computer-implemented method of claim 2, further comprising:

receiving, by the one or more computer systems, a request for information; and

applying the supervised labeling model to the request for information to determine a label for the request.

4. The computer-implemented method of claim 3, further comprising:

determining that a confidence level for a relevance of information in the cross-channel cluster is less than a threshold level; and

generating a message that identifies information in the cross-channel cluster that is likely to address the request for information.

5. The computer-implemented method of claim 3, further comprising:

identifying that the cross-channel cluster includes information that is relevant to the request for information, when a confidence level for a relevance of the relevant information in the cross-channel cluster exceeds a threshold confidence level;

ranking the relevant information associated with the cross-channel cluster; and

generating, based on the ranking, a response to the request, with the response comprising at least a portion of the ranked information.

6. The computer-implemented method of claim 3, further comprising:

identifying a correspondence between the label for the request and a label for the cross-channel cluster;

ranking information associated with the cross-channel cluster; and

generating, based on the ranking, a response to the request, with the response comprising at least a portion of the ranked information.

7. The computer-implemented method of claim 1, wherein the first type of channel comprises one of an electronic mail channel, an online forum channel, an online help center channel, or an instant messaging channel, and wherein the second type of channel comprises one of an electronic mail channel, an online forum channel, an online help center channel, or an instant messaging channel.

8. The computer-implemented method of claim 1, further comprising:

generating a link between the information associated with the first type of channel in the plurality of channels and the information associated with the second type of channel in the plurality of channels.

9. The computer-implemented method of claim 1, further comprising:

in response to a request,

selecting, from the cross-channel cluster, information items relevant to the request.

10. A computer-implemented method comprising:

receiving, by one or more computer systems, first information from a first type of channel in a plurality of channels and second information from a second type of channel in the plurality of channels;

generating a first cluster of information items for the first type of channel in the plurality of channels, with the information items in the first cluster associated with similar topics;

applying a supervised labeling model to generate a first label for the first cluster, with the first label indicative of the topics associated with the items of information in the first cluster;

generating a second cluster of information items for the second type of channel in the plurality of channels, with the information items in the second cluster associated with similar topics, with the first type of channel differing from the second type of channel;

applying the supervised labeling model to generate a second label for the second cluster, with the second label indicative of the topics associated with the items of information in the second cluster;

identifying a correspondence between the first label for the first cluster and the second label for the second cluster;

merging the first cluster of information with the second cluster of information;

generating, based on merging, a cross-channel cluster of information wherein the cross-channel cluster comprises a cluster of information that is received from at least two different types of channels; and

in response to a request, selecting information items relevant to the request using the merged first and second clusters.

11. A computer-implemented method comprising:

accessing a first cluster of a first plurality of resources belonging to a first type of channel in a plurality of channels, with each resource in the first cluster associated with a topic, with the first type of channel comprising one or more an electronic mail (e-mail) channel, an online forum channel, an online help center channel, an instant messaging channel, and a knowledge base channel, and with a resource comprising one or more of text information, a document, a message, a web forum request, and a support inquiry;

accessing a second cluster of a second plurality of resources belonging to a second type of channel in the plurality of channels, with each resource in the second cluster associated with a topic;

for each resource in the first plurality of resources and in the second plurality of resources;

applying an unsupervised clustering model to the resource; and

assigning, based on the applying, the resource to a cross-channel cluster from a plurality of cross-channel clusters, with the cross-channel cluster including other resources associated with topics that are similar to a topic associated with the assigned resource, and with the cross-channel cluster comprising a cluster of information that is received from at least two different types of channels;

receiving a request for resources associated with a topic;

determining a correspondence between (a) the topic associated with the request for resources, and (b) at least one of the cross-channel clusters;

selecting one or more resources included in a cross-channel cluster that corresponds to the topic associated with the request; and

sending the selected one or more resources to a computing device initiating the request.

12. One or more machine-readable media configured to store instructions that are executable by one or more processing devices to perform operations comprising:

receiving first information from a first type of channel and second information from a second type of channel, with the first type of channel differing from the second type of channel;

merging the first information received from the first type of channel with the second information received from the second type of channel that differs from the first type of channel;

applying an unsupervised clustering model to the merged information; and

generating, based on results of the applying, a cross-channel cluster, the cross-channel cluster comprising (i) a portion of the first information received from the first type of channel associated with a subject matter, and (ii) a portion of the second information received from the second type of channel that differs from the first type of channel associated with the subject matter, wherein the cross-channel cluster comprises a cluster of information that is received from at least two different types of channels.

13. The one or more machine-readable of claim 12, wherein the operations further comprise:

applying a supervised labeling model to identify a label for the cross-channel cluster.

14. The one or more machine-readable of claim 13, wherein the operations further comprise:

receiving, by the one or more computer systems, a request for information; and

applying the supervised labeling model to the request for information to determine a label for the request.

15. The one or more machine-readable of claim 14, wherein the operations further comprise:

determining that a confidence level for a relevance of information in the cross-channel cluster is less than a threshold level; and

generating a message that identifies information in the cross-channel cluster that is likely to address the request for information.

16. An electronic system, comprising:

one or more processing devices; and

one or more machine-readable media configured to store instructions that are executable by the one or more processing devices to perform operations comprising;

receiving first information from a first type of channel and second information from a second type of channel, with the first type of channel differing from the second type of channel;

merging the first information received from the first type of channel with the second information received from the second type of channel that differs from the first type of channel;

applying an unsupervised clustering model to the merged information; and

generating, based on results of the applying, a cross-channel cluster, the cross-channel cluster comprising (i) a portion of the first information received from the first type of channel associated with a subject matter, and (ii) a portion of the second information received from the second type of channel that differs from the first type of channel associated with the subject matter, wherein the cross-channel cluster comprises a cluster of information that is received from at least two different types of channels.

17. The electronic system of claim 16, wherein the operations further comprise:

applying a supervised labeling model to identify a label for the cross-channel cluster.

18. The electronic system of claim 17, wherein the operations further comprise:

receiving, by the one or more computer systems, a request for information; and

applying the supervised labeling model to the request for information to determine a label for the request.

19. The electronic system of claim 18, wherein the operations further comprise:

determining that a confidence level for a relevance of information in the cross-channel cluster is less than a threshold level; and

generating a message that identifies information in the cross-channel cluster that is likely to address the request for information.

20. One or more machine-readable media configured to store instructions that are executable by one or more processing devices to perform operations comprising:

sending a request for information; and

receiving one or more resources included in a cross-channel cluster, with each resource corresponding to a topic associated with the request,

wherein the cross-channel cluster comprises (i) a portion of first information from a first type of channel, the portion of first information associated with the topic, and (ii) a portion of second information from a second type of channel that differs from the first type of channel, the portion of second information associated with the topic, wherein the cross-channel cluster comprises a cluster of information that is received from at least two different types of channels.