System and Method for Identification of Intent Segment(s) in Caller-Agent Conversations

• US 20140244249A1
• Filed: 02/28/2013
• Published: 08/28/2014
• Est. Priority Date: 02/28/2013
• Status: Active Grant

• Current Assignee: International Business Machines Corporation, Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation, Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Verma, Ashish, Marcus, Jeffrey N., Mohamed, Shajith Ikbal, Church, Kenneth W., Ghosh, Prasanta
• Granted Patent: US 10,354,677 B2
• Time in Patent Office: Days
• US Class Current: 704/233
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 2015/088 Word spotting
• CPC Class Codes: G10L 25/63 for estimating an emotional state

Abstract

Identification of an intent of a conversation can be useful for real-time or post-processing purposes. According to example embodiments, a method, and corresponding apparatus of identifying at least one intent-bearing utterance in a conversation, comprises determining at least one feature for each utterance among a subset of utterances of the conversation; classifying each utterance among the subset of utterances, using a classifier, as an intent classification or a non-intent classification based at least in part on a subset of the at least one determined feature; and selecting at least one utterance, with intent classification, as an intent-bearing utterance based at least in part on classification results by the classifier. Through identification of an intent bearing utterance, a call center for example, can provide improved service for callers through, for example, more effective directing of a call to a live agent.

Claims

1. A method of identifying at least one intent-bearing utterance in a conversation, the method comprising:

determining at least one feature for each utterance among a subset of utterances of the conversation;

classifying each utterance among the subset of utterances, using a classifier, as an intent classification or a non-intent classification based at least in part on a subset of the at least one determined feature; and

selecting at least one utterance, with intent classification, as an intent-bearing utterance based at least in part on classification results by the classifier.

2. The method according to claim 1, wherein selecting at least one utterance of the subset of utterances includes:

assigning a score for each utterance of the subset of utterances based at least in part on the classification results; and

selecting the at least one utterance based at least in part on the score assigned for said each utterance of the subset of utterances.

3. The method according to claim 1, wherein classifying each utterance of the subset of utterances includes:

generating a plurality of state sequences, each state sequence describing a sequence of classifications associated with the subset of utterances.

4. The method according to claim 3, wherein selecting at least one utterance includes:

selecting the at least one utterance based on a highest occurrence of intent classification among the plurality of state sequences generated.

5. The method according to claim 1, wherein the conversation is between a customer service agent and a customer.

6. The method according to claim 1, wherein the at least one feature includes at least one of the following:

at least one word associated with the utterance;

identity of a participant associated with utterance; and

a relative position of the utterance within the call.

7. The method according to claim 6, wherein the at least one word associated with the utterance includes a least one of:

a word or expression frequently used, by a customer service agent, prior to an intent-bearing utterance;

a word or expression frequently used, by a customer service agent, following an intent-bearing utterance; and

a word or expression frequently used, by a customer or a customer service agent, within an intent-bearing utterance.

8. The method according to claim 7, wherein the word or expression frequently used prior to, within, or following an intent bearing utterance being provided to the classifier manually through rules or automatically identified in a data driven manner.

9. An apparatus for identifying at least one intent-bearing utterance in a conversation, the apparatus comprising:

a memory with computer code instructions stored thereon; and

a processor,

the memory, with the computer code instructions, and the processor being configured to;

determine at least one feature for each utterance among a subset of utterances of the conversation;

classify each utterance among the subset of utterances, using a classifier, as an intent classification or a non-intent classification based at least in part on a subset of the at least one determined feature; and

select at least one utterance, with intent classification, as an intent-bearing utterance based at least in part on classification results by the classifier.

10. The apparatus according to claim 9, wherein in selecting at least one utterance of the subset of utterances, the memory with the computer code instructions and the processor are configured to:

assign a score for each utterance of the subset of utterances based at least in part on the classification results; and

select the at least one utterance based at least in part on the score assigned for said each utterance of the subset of utterances.

11. The apparatus according to claim 9, wherein in classifying each utterance of the subset of utterances, the memory with the computer code instructions and the processor are configured to:

generate a plurality of state sequences, each state sequence describing a sequence of classifications associated with the subset of utterances.

12. The apparatus according to claim 11, wherein in selecting at least one utterance the memory with the computer code instructions and the processor are configured to:

select the at least one utterance based on a highest occurrence of intent classification among the plurality of state sequences generated.

13. The apparatus according to claim 9, wherein the conversation is between a customer service agent and a customer.

14. The apparatus according to claim 10, wherein the at least one feature includes at least one of the following:

at least one word associated with the utterance;

identity of a participant associated with utterance; and

a relative position of the utterance within the call.

15. The apparatus according to claim 14, wherein the at least one word associated with the utterance includes at least one of:

a word or expression frequently used by customer service agent prior to an intent-bearing utterance;

a word or expression frequently used by customer service agent following an intent-bearing utterance; and

a word or expression frequently used, by a customer or a customer service agent, within an intent-bearing utterance.

16. The apparatus according to claim 15, wherein the word or expression frequently used prior to, within, or following an intent bearing utterance being provided to the classifier manually through rules or automatically identified in a data driven manner.

17. A non-transitory computer-readable medium for identifying at least one intent-bearing utterance in a conversation, the non-transitory computer-readable medium includes computer code stored thereon, the computer code when executed by a processor causes an apparatus to perform at least the following:

determining at least one feature for each utterance among a subset of utterances of the conversation;

classifying each utterance among the subset of utterances, using a classifier, as an intent classification or a non-intent classification based at least in part on a subset of the at least one determined feature; and

selecting at least one utterance, with intent classification, as an intent-bearing utterance based at least in part on classification results by the classifier.

18. The non-transitory computer-readable medium according to claim 17, wherein in selecting at least one utterance of the subset of utterances, the computer code when executed by a processor causes an apparatus to further perform:

assigning a score for each utterance of the subset of utterances based at least in part on the classification results; and

selecting the at least one utterance based at least in part on the score assigned for said each utterance of the subset of utterances.

19. The non-transitory computer-readable medium according to claim 18, wherein in classifying each utterance of the subset of utterances, the computer code when executed by a processor causes an apparatus to further generate a plurality of state sequences, each state sequence describing a sequence of classifications associated with the subset of utterances, and wherein in selecting at least one utterance, the computer code when executed by a processor causes the apparatus to select the at least one utterance based on highest occurrence of intent classification among the plurality of state sequences generated.