Transcription generation technique selection

Patent metadata

Field	Value
Publication number	US-10867609-B2
Application number	US-201815984081-A
Country	US
Kind code	B2
Filing date	May 18, 2018
Priority date	May 18, 2018
Publication date	Dec 15, 2020
Grant date	Dec 15, 2020

Abstract

According to one or more aspects of the present disclosure, operations related to selecting a transcription generation technique may be disclosed. In some embodiments, the operations may include obtaining multiple user ratings that each correspond to a different one of multiple transcriptions. Each transcription may be obtained using a first transcription generation technique and may correspond to a different one of multiple communication sessions. The operations may further include selecting, for a subsequent communication session that occurs after the multiple communication sessions, a second transcription generation technique based on the user ratings. In addition, the operations may include providing the subsequent transcription to a device during the subsequent communication session.

First claim

What is claimed is:

1. A method comprising:
obtaining, using a first transcription generation technique, a first transcription of a first communication session between a first device associated with a first user and a second device associated with a second user;
providing the first transcription to the first device during the first communication session;
obtaining a first user rating of the first transcription, the first user rating indicating feedback from the first user regarding the first transcription of the first communication session;
after termination of the first communication session, obtaining, using the first transcription generation technique, a second transcription of a second communication session between the first device and a third device associated with a third user;
providing the second transcription to the first device during the second communication session;
obtaining a second user rating of the second transcription, the second user rating indicating feedback from the first user regarding the second transcription of the second communication session;
after obtaining the first and second user ratings and after termination of the second communication session, selecting, based on the first and second user ratings, a second transcription generation technique for a third communication session between the first device and a fourth device; and
providing, to the first device during the third communication session, a third transcription that is obtained using the second transcription generation technique.

2. The method of claim 1 , wherein:
the first transcription generation technique includes using a fully machine based automatic speech recognition system to generate the first transcription and the second transcription; and
the second transcription generation technique includes using a re-voicing speech recognition system to generate the third transcription.

3. The method of claim 1 , wherein:
the first transcription generation technique includes using a first fully machine based automatic speech recognition system to generate the first transcription and the second transcription; and
the second transcription generation technique includes using a second fully machine based automatic speech recognition system to generate the third transcription.

4. The method of claim 1 , wherein obtaining the first user rating includes:
directing presentation of a request for the first user rating on the first device after termination of the first communication session; and
obtaining the first user rating from the first device based on a response to the request.

5. The method of claim 1 , wherein selecting the second transcription generation technique includes:
determining a number of negative user ratings of a plurality of user ratings from the first user, which include the first user rating and the second user rating, that meet a negative rating threshold; and
selecting the second transcription generation technique based on the number of negative user ratings meeting a number threshold.

6. The method of claim 1 , wherein selecting the second transcription generation technique includes:
determining a ratio of negative user ratings that meet a negative rating threshold with respect to a plurality of user ratings from the first user that include the first user rating and the second user rating; and
selecting the second transcription generation technique based on the ratio of negative user ratings meeting a ratio threshold.

7. The method of claim 1 , wherein selecting the second transcription generation technique includes:
determining an average user rating of a plurality of user ratings from the first user that include the first user rating and the second user rating; and
selecting the second transcription generation technique based on the average user rating meeting a negative rating threshold.

8. At least one non-transitory computer-readable media configured to store one or more instructions that in response to being executed by at least one computing system cause performance of the method of claim 1 .

9. A method comprising:
obtaining a first user rating of a first transcription of a first communication session between a first device associated with a first user and a second device associated with a second user, the first user rating indicating feedback from the first user regarding the first transcription and the first transcription obtained using a first transcription generation technique;
after termination of the first communication session, obtaining a second user rating of a second transcription of a second communication session between the first device and a third device associated with a third user, the second user rating indicating feedback from the first user regarding the second transcription and the second transcription obtained using the first transcription generation technique;
selecting, based on the first and second user ratings, a second transcription generation technique for a third communication session between the first device and a fourth device;
obtaining, during the third communication session, a third transcription that is obtained using the second transcription generation technique; and
providing the third transcription to the first device during the third communication session.

10. The method of claim 9 , wherein:
the first transcription generation technique includes using a fully machine based automatic speech recognition system to generate the first transcription and the second transcription; and
the second transcription generation technique includes using a re-voicing speech recognition system to generate the third transcription.

11. The method of claim 9 , wherein:
the first transcription generation technique includes using a first fully machine based automatic speech recognition system to generate the first transcription and the second transcription; and
the second transcription generation technique includes using a second fully machine based automatic speech recognition system to generate the third transcription.

12. The method of claim 9 , wherein obtaining the first user rating includes:
directing presentation of a request for the first user rating on the first device after termination of the first communication session; and
obtaining the first user rating from the first device based on a response to the request.

13. The method of claim 9 , wherein selecting the second transcription generation technique includes:
determining a number of negative user ratings of a plurality of user ratings from the first user, which include the first user rating and the second user rating, that meet a negative rating threshold; and
selecting the second transcription generation technique based on the number of negative user ratings meeting a number threshold.

14. The method of claim 9 , wherein selecting the second transcription generation technique includes:
determining a ratio of negative user ratings that meet a negative rating threshold with respect to a plurality of user ratings from the first user that include the first user rating and the second user rating; and
selecting the second transcription generation technique based on the ratio of negative user ratings meeting a ratio threshold.

15. The method of claim 9 , wherein selecting the second transcription generation technique includes:
determining an average user rating of a plurality of user ratings from the first user that include the first user rating and the second user rating; and
selecting the second transcription generation technique based on the average user rating meeting a negative rati

Assignees

• Sorenson Ip Holdings Llc

Inventors

• Holm Michael

Classifications

• G10L15/30

  Distributed recognition, e.g. in client-server systems, for mobile phones or network applications · CPC title

• G10L15/28

  Constructional details of speech recognition systems · CPC title

• G10L15/18

  using natural language modelling · CPC title

• G10L15/16

  using artificial neural networks · CPC title

• G10L15/14

  using statistical models, e.g. Hidden Markov Models [HMMs] (G10L15/18 takes precedence) · CPC title