Systems and methods for speech analytics and phrase spotting using phoneme sequences

Patent metadata

Field	Value
Publication number	US-2016019882-A1
Application number	US-201414332115-A
Country	US
Kind code	A1
Filing date	Jul 15, 2014
Priority date	Jul 15, 2014
Publication date	Jan 21, 2016
Grant date	—

Abstract

A contact center system can receive audio messages. The system can review audio messages by identifying phoneme strings within the audio messages associated with a characteristic. A phoneme can be a component of spoken language. Identified phoneme strings are used to analyze subsequent audio messages to determine the presence of the characteristic without requiring human analysis. Thus, the identification of phoneme strings then can be used to determine a characteristic of audio messages without transcribing the messages.

First claim

What is claimed is:

1 . A method for determining a characteristic in an audio message, the method comprising:
determining a phoneme in the audio message having a predetermined characteristic; identifying a first phoneme string in the audio message, wherein the phoneme string includes the phoneme, and wherein the first phoneme string is associated with the predetermined characteristic; and based on the identification of the first phoneme string, determining that the first phoneme string indicates the characteristic.

2 . The method as defined in claim 1 , further comprising:
receiving a second message; and identifying the first phoneme string within the second message.

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

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

5 . The method as defined in claim 4 , wherein the characteristic may be a sentiment, and wherein the sentiment may be positive or negative.

6 . The method as defined in claim 4 , further comprising:
receiving a new set of audio messages; identifying a second phoneme string in the new set of audio messages, wherein the second phoneme string includes a second phoneme, and wherein the second phoneme string is associated with a second characteristic; comparing the second phoneme string in the new set of audio messages with at least two audio messages in an old set of audio messages; based on the comparison, determining that the second phoneme string is absent from the old set of audio messages; and determining that the second characteristic is a new topic present in the new set of audio messages.

7 . The method as defined in claim 4 , wherein the confidence score is a probability, the method further comprising:
determining if the confidence score reaches or crosses a predetermined threshold; and if the confidence score reaches or crosses the predetermined threshold, signifying that the first phoneme string indicates the characteristic.

8 . The method as defined in claim 5 , further comprising:
if the confidence score does not reach or cross the predetermined threshold, receiving a third message; re-calculating the confidence score using analysis associated with the third message.

9 . The method as defined in claim 6 , wherein each phoneme string includes two or more phonemes associated therewith.

10 . The method as defined in claim 7 , wherein each message includes two or more phonemes.

11 . The method as defined in claim 8 , wherein the phonemes are associated with the English language.

12 . The method as defined in claim 9 , further comprising:
analyzing a known negative/positive message; identifying phoneme strings; and storing the phoneme strings.

13 . A computer readable medium having stored thereon processor executable instructions that cause a computing system to execute a method, the instructions comprising:
instructions to receive a first message; instructions to identify a phoneme within the first message; instructions to output an index file listing all phonemes within the first message instructions to analyze the first message for a characteristic; instructions to determine a first phoneme string in the index file that is associated with the characteristic; instructions to store the phoneme string; instructions to retrieve the phoneme string; instructions to identify the phoneme string, associated with the characteristic, in a second message; and based on the identification of the phoneme string, instructions to determine that the second message has the characteristic.

14 . The computer readable medium as defined in claim 11 , further comprising instructions to determine statistical information about the phoneme string, wherein the statistical information includes a confidence score that the phoneme string indicates a characteristic.

15 . The computer readable medium as defined in claim 12 , further comprising:
instructions to determine if the confidence score reaches or crosses a predetermined threshold; if the confidence score reaches or crosses the predetermined threshold, instructions to signify that the phoneme string indicates a known characteristic; if the confidence score does not reach or cross the predetermined threshold, instructions to receive a third message; and instructions to re-calculate the confidence score using analysis associated with the third message.

16 . The computer readable medium as defined in claim 13 , wherein each phoneme string includes two or more phonemes associated therewith.

17 . The computer readable medium as defined in claim 14 , wherein each message includes two or more phonemes.

18 . A communication system comprising:
a dialog system, the dialog system operable to determine an agent routing for an audio message, wherein the dialog system compromises:
a processing component that is operable to receive and analyze the audio message, wherein the processing component compromises:
a phoneme identifier operable to:
receive a first message;
identify two or more phoneme within the first message;
output an index file listing the two or more phonemes within the first message, wherein the index file lists the two or more phonemes in an order identified in the first message;

a phoneme string identifier in communication with the phoneme identifier, wherein the phoneme string identifier is operable to analyze the first message for a phoneme string;
a parser in communication with the phoneme string identifier, wherein the parser is operable store the phoneme string;
a message characteristic identifier in communication with the parser, wherein the characteristic identifier is operable to:
retrieve the phoneme string associated with a known characteristic;
identify the phoneme string in a second message; and
based on the identification of the phoneme string, determine that the second message has a known characteristic.

19 . The communication system as defined in claim 16 , further comprising a statistics analyzer in communication with the parser, wherein the statistics analyzer is operable to determine statistical information about the phoneme string, wherein the statistical information includes a confidence score that the phoneme string indicates a characteristic.

20 . The communication system as defined in claim 17 , wherein the statistics analyzer is further operable to:
determine if the confidence score reaches or crosses a predetermined threshold; if the confidence score reaches or crosses the predetermined threshold, signify that the phoneme string indicates a known characteristic; if the confidence score does not reach or cross the predetermined threshold, receive a third message; and instructions to re-calculate the confidence score use analysis associated with the third message.

Assignees

• Avaya Inc

Inventors

• Matula Valentine C
• Shaffer Shmuel

Classifications

• G10L2015/025

  Phonemes, fenemes or fenones being the recognition units · CPC title

• G10L25/51

  for comparison or discrimination · CPC title

• G10L15/187 Primary

  Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams · CPC title

• G10L15/02 Primary

  Feature extraction for speech recognition; Selection of recognition unit · CPC title