Granted Patent B2

US 11,289,077 · App. 14/332,115 · Granted Mar 29, 2022

Inventors: Valentine C. Matula (Granville, OH); Shmuel Shaffer (Palo Alto, CA)

Assignee: Avaya Inc.

G10L15/187G10L25/51G10L2015/025

Abstract

Claims (68)

1. A computer-implemented method for determining a characteristic in audio messages, the method comprising:

receiving, by a processor, a first audio message via a network;

partitioning, by the processor, the first audio message into two or more frames;

identifying, by the processor, a first phoneme and a second phoneme in the two or more frames;

producing, by the processor, a first index file by concatenating the two or more frames, wherein the first index file comprises a concatenation of the first and second phonemes;

identifying, by the processor, a first phoneme string in the first index file by comparing the first index file with one or more existing index files, wherein the first phoneme string is associated with a first characteristic;

based on identifying the first phoneme string in the first index file, determining, by the processor, that the first phoneme string indicates the first characteristic;

receiving, by the processor, a second audio message from a website via the network;

identifying, by the processor, the first phoneme string in the second audio message; and

in response to identifying the first phoneme string in the second audio message, determining, by the processor, that the second audio message indicates the first characteristic.

2. The method as defined in claim 1 , further comprising determining statistical information about the first phoneme string.

3. The method as defined in claim 2 , wherein the statistical information includes a confidence score indicating a degree of confidence that the first phoneme string indicates the first characteristic.

4. The method as defined in claim 3 , wherein the first characteristic is associated with one of a positive sentiment and a negative sentiment.

5. The method as defined in claim 3 , further comprising:

determining the confidence score does not reach or cross a predetermined threshold;

receiving a third message; and

in response to determining the confidence score does not reach or cross the predetermined threshold and in response to receiving the third message, re-calculating the confidence score using analysis associated with the third message.

6. The method as defined in claim 1 , wherein the first and second phonemes are associated with the English language.

7. The method as defined in claim 3 , further comprising:

receiving, by the processor, a first set of audio messages;

identifying, by the processor, a second phoneme string in the first set of audio messages, wherein the second phoneme string is associated with a second characteristic;

comparing, by the processor, the second phoneme string in the first set of audio messages with at least two audio messages in a second set of audio messages;

based on the comparing, determining, by the processor, the second phoneme string is absent from the second set of audio messages; and

determining, by the processor, the second characteristic is a new topic present in the first set of audio messages.

8. The method as defined in claim 7 , wherein each of the first and second phoneme strings includes two or more phonemes.

9. The method as defined in claim 8 , further comprising:

analyzing a known negative message; and

identifying phoneme strings in the known negative message.

10. The method as defined in claim 3 , wherein the confidence score is a probability, the method further comprising:

determining the confidence score reaches or crosses a predetermined threshold; and

based on determining the confidence score reaches or crosses the predetermined threshold, signifying the first phoneme string indicates the first characteristic.

11. The method as defined in claim 10 , wherein each of the first and second audio messages includes two or more phonemes.

12. A non-transitory computer-readable storage medium having stored thereon computer-readable processor-executable instructions configured to cause one or more processors of a computing system to execute a method, the method comprising:

receiving a first audio message via a network;

partitioning the first audio message into two or more frames;

identifying a first phoneme and a second phoneme in the two or more frames;

producing a first index file by concatenating the two or more frames, wherein the first index file comprises a concatenation of the first and second phonemes;

identifying a first phoneme string in the first index file by comparing the first index file with one or more existing index files, wherein the first phoneme string is associated with a first characteristic;

based on identifying the first phoneme string in the first index file, determining, by the processor, that the first phoneme string indicates the first characteristic;

receiving, by the processor, a second audio message from a website via the network;

identifying the first phoneme string, associated with the first characteristic, in the second audio message;

identifying, by the processor, the first phoneme string in the second audio message; and

in response to identifying the first phoneme string in the second audio message, determining that the second audio message indicates the first characteristic.

13. The computer readable storage medium as defined in claim 12 , further comprising instructions to determine statistical information about the first phoneme string, wherein the statistical information includes a confidence score indicating a degree of confidence that the first phoneme string indicates the first characteristic.

14. The computer readable storage medium as defined in claim 13 , the method further comprising:

determining the confidence score reaches or crosses a predetermined threshold;

based on determining the confidence score reaches or crosses the predetermined threshold, signifying the first phoneme string indicates the first characteristic.

15. The computer readable storage medium as defined in claim 13 , wherein each of the first audio message and the second audio message includes two or more phonemes.

16. The computer readable storage medium as defined in claim 12 , wherein the first phoneme string comprises two or more phonemes.

17. A communication system comprising:

a dialog system operable to determine an agent routing for an audio message, wherein the dialog system comprises a processing component operable to receive and analyze audio messages, wherein the processing component comprises:

a phoneme identifier operable to:

receive a first audio message via a network;

partition the first audio message into two or more frames;

identify a first phoneme and a second phoneme in the two or more frames; and

produce a first index file by concatenating the two or more frames, wherein the first index file comprises a concatenation of the first and second phonemes;

a phoneme string identifier in communication with the phoneme identifier, wherein the phoneme string identifier is operable to:

identify a first phoneme string in the first index file by comparing the first index file with one or more existing index files, wherein the first phoneme string is associated with a first characteristic; and

based on identifying the first phoneme string in the first index file, determine that the first phoneme string indicates the first characteristic; and

a message characteristic identifier operable to:

receive a second audio message from a website via the network;

identify the first phoneme string in the second audio message; and

in response to identifying the first phoneme string in the second audio message, determine that the second audio message indicates the first characteristic.

18. The communication system as defined in claim 17 , further comprising a statistics analyzer operable to determine statistical information about the first phoneme string, wherein the statistical information includes a confidence score indicating a degree of confidence that the first phoneme string indicates the first characteristic.

19. The communication system as defined in claim 18 , wherein the statistics analyzer is further operable to:

determine the confidence score reaches or crosses a predetermined threshold; and

based on determining the confidence score reaches or crosses the predetermined threshold, signify that the first phoneme string indicates the first characteristic.

20. The communication system as defined in claim 18 , wherein the first characteristic is one of a positive sentiment and a negative sentiment.

Assignments (17)

Continuity (1)