Semantic clustering and user interfaces

• US 9,275,042 B2
• Filed: 01/24/2014
• Issued: 03/01/2016
• Est. Priority Date: 03/26/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: VirtuOz SA (Microsoft Corporation)
• Inventors: Larcheveque, Jean-Marie Henri Daniel, Powers, Elizabeth Ireland, Recksiek, Freya Kate, Teodosiu, Dan
• Primary Examiner(s): PULLIAS, JESSE SCOTT
• Application Number:
  US14/163,744
• Publication Number:
  US 20140207441A1
• Time in Patent Office: 767 Days
• Field of Search: 704/1, 704/9, 704/10, 704/257
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

Semantic clustering techniques are described. In various implementations, a conversational agent is configured to perform semantic clustering of a corpus of user utterances. Semantic clustering may be used to provide a variety of functionality, such as to group a corpus of utterances into semantic clusters in which each cluster pertains to a similar topic. These clusters may then be leveraged to identify topics and assess their relative importance, as for example to prioritize topics whose handling by the conversation agent should be improved. A variety of utterances may be processed using these techniques, such as spoken words, textual descriptions entered via live chat, instant messaging, a website interface, email, SMS, a social network, a blogging or micro-blogging interface, and so on.

Claims

1. A method implemented by one or more computer processors configured to perform semantic clustering, the method comprising:

computationally performing a linguistics analysis of a corpus of user utterances to identify semantic graphs that match respective user utterances in the corpus;

computationally clustering the semantic graphs of the utterances in the corpus based on user intent to form one or more semantic clusters; and

attaching additional semantic graphs of user utterances to the one or more semantic clusters based on the user intent.

2. The method as claimed in claim 1, wherein the corpus is a first corpus, and wherein the method further comprises:

computationally processing through a linguistics analysis of a second corpus of user utterances to identify semantic graphs that match respective user utterances in the second corpus, wherein the additional attached semantic graphs of user utterances are from the second corpus.

3. The method as claimed in claim 2, wherein the one or more semantic clusters correspond to topics that occur in the first corpus but not in the second corpus.

4. The method as claimed in claim 3, wherein at least one of said topics is expressed by at least two semantic graphs formed by linguistic analysis and from first and second sets of semantic clusters using different words and grammatical structures.

5. The method as claimed in claim 2, wherein the first and second corpa correspond to different time intervals.

6. The method as claimed in claim 2, wherein user utterances from at least the first or second corpa are derived from a natural language dialog with an executing conversational agent.

7. The method as claimed in claim 6 wherein the executing conversational agent is configured to leverage the one or more semantic clusters in order to respond to interaction with the user.

8. The method as claimed in claim 2, wherein user utterances from at least the first or second corpa are derived from voice input provided by a client device.

9. The method as claimed in claim 2, wherein user utterances from at least one of the first or second corpora are derived from an instant messaging conversation, or a feed from a short messaging service (SMS), social networking, blogging, or micro-blogging service.

10. A computing device comprising:

a processor;

a computer storage operationally coupled to the processor and configured to store a corpus of user utterances received by a conversational agent, configured to be processed computationally by the processor, through an interactive natural language dialog with a user;

the conversational agent further configured to perform computationally a linguistics analysis of the corpus of user utterances to identify semantic graphs that match respective user utterances in the corpus;

the executing conversational agent still further configured to;

cluster computationally the semantic graphs of the utterances in the first corpus based on user intent to form one or more semantic clusters; and

attach additional semantic graphs of user utterances to the one or more semantic clusters based on the user intent.

11. The computing device as claimed in claim 10 wherein the corpus is a first corpus;

and wherein the method further includes the conversational agent being further configured to;

computationally process though a linguistics analysis of a second corpus of user utterances to identify semantic graphs that match respective user utterances in the second corpus, wherein the additional attached semantic graphs of user utterances are from the second corpus.

12. The computing device as claimed in claim 11, wherein the one or more semantic clusters correspond to topics that occur in the first corpus but not in the second corpus.

13. The computing device as claimed in claim 11, wherein at least one of said topics is expressed by at least two semantic graphs formed by linguistic analysis and from first and second sets of semantic clusters using different words and grammatical structures.

14. The computing device as claimed in claim 11, wherein the first and second corpa correspond to different time intervals.

15. The computing device as claimed in claim 11, wherein user utterances from at least the first or second corpa are derived from a natural language dialog with the executing conversational agent.

16. The computing device as claimed in claim 11, wherein user utterances from at least the first or second corpa are derived from voice input received via a mobile phone device.

17. The computing device as claimed in claim 11, wherein user utterances from at least one of the first or second corpora are derived from an instant messaging conversation, or a feed from a short messaging service (SMS), social networking, blogging, or micro-blogging service.

18. The computing device as claimed in claim 10 wherein the executing conversation agent is further configured to leverage the one or more semantic clusters in order to respond to interaction with the user.

19. A computer program product stored on a non-transitory computer useable medium having computer readable instructions, which, when executed by one or more computer processors, cause the one or more computer processers to execute a conversation agent configured to perform semantic clustering by:

receiving a subject user utterance through a natural language dialog with the conversation agent;

forming an intent graph pattern of that subject user utterance;

comparing the intent graph pattern of the subject user utterance with one or more clusters of user utterances to identify a semantic graph of a user utterance in the cluster that matches the intent graph pattern of the subject user utterance;

responding to the subject user utterance based at least in part on the semantic graph in the cluster that matches the intent graph pattern of the subject user utterance; and

attaching a semantic graph of the subject user utterance to the cluster based on the user intent.

20. The computer program product as claimed in claim 19, wherein the subject user utterance is part of a corpus of a plurality of user utterances, which are derived from a natural language dialog with the conversation agent.