Voice recognition server and control method thereof

Patent metadata

Field	Value
Publication number	US-10510342-B2
Application number	US-201615063872-A
Country	US
Kind code	B2
Filing date	Mar 8, 2016
Priority date	Sep 16, 2015
Publication date	Dec 17, 2019
Grant date	Dec 17, 2019

Abstract

Provided herein is a voice recognition server and a control method thereof, the method including determining an index value for each of a plurality of training texts; setting a group for each of the plurality of training texts based on the index values of the plurality of training texts, and matching a function corresponding to each group and storing the matched results; in response to receiving a user's uttered voice from a user terminal apparatus, determining an index value from the received uttered voice; and searching a group corresponding to the index value determined from the received uttered voice, and performing the function corresponding to the uttered voice, thereby providing a voice recognition result of a variety of user's uttered voices suitable to the user's intentions.

First claim

What is claimed is:

1. A method for controlling a voice recognition server, the method comprising:
identifying an index value which is a value that is located at one point within a continuous space for each of a plurality of training texts by a processor;
setting a group for each of the plurality of training texts based on a similarity of the index value and a distance between the index value and locations of a predefined plurality of groups by the processor, and wherein the setting the group comprises setting an average value of the different index values included in the group as the index value of the group;
matching a function corresponding to each group by a processor;
storing the matched results;
identifying an index value from a received uttered voice in response to receiving a user's uttered voice from a user terminal apparatus by a processor;
searching for a group corresponding to the index value identified from the received uttered voice by comparing the index value of the group with index value identified from the received uttered voice by a processor; and
performing the function corresponding to the uttered voice based on the group corresponding to the index value by a processor.

2. The method according to claim 1 , wherein the identifying an index value for each training text comprises: dividing the training text into meaning units; identifying the index value having the meaning units as scalar or vector values; and tagging the identified index value to each of the meaning units.

3. The method according to claim 2 , wherein the index value is to be identified using a Skip-Gram method.

4. The method according to claim 2 , further comprising:
setting a function corresponding to the set group;
tagging the function to the group; and
storing the tagged result.

5. The method according to claim 4 ,
wherein the setting a group further includes setting an index value closest to the average value as the index value of the group.

6. The method according to claim 5 ,
wherein the setting a function includes setting at least one of: a command processing function, an intention processing function and a sentiment processing function as the function corresponding to the group, based on a received command.

7. The method according to claim 6 ,
wherein the identifying an index value from the received uttered voice comprises: dividing a text recognized from the uttered voice into meaning units; and identifying the index value by expressing the meaning units divided from the text as scalar or vector values.

8. The method according to claim 7 ,
wherein the performing the function comprises: selecting a group that corresponds to the identified index value among the pre-stored groups; and transmitting a voice recognition result corresponding to the function of the selected group to the user terminal apparatus.

9. The method according to claim 8 ,
wherein the transmitting a voice recognition result includes, in response to the function of the selected group being the command processing function, transmitting a control command for controlling an operation of the user terminal apparatus to the user terminal apparatus, and
in response to the function of the selected group being the intention processing function or the sentiment processing function, transmitting at least one of: a response message, control message, and search result of the uttered voice, to the user terminal apparatus.

10. A voice recognition server comprising:
memory;
a processor configured to:
identify an index value, which is a value that is located at one point within a continuous space, of each of a plurality of training texts,
set a group for each of the plurality of training texts based on a similarity of the index value and a distance between the index value and locations of a predefined plurality of groups, and wherein the set the group comprises the processor configured to set an average value of the different index values included in the group as the index value of the group;
match a function corresponding to each group,
control to memory to store the matched results,
in response to receiving a user's uttered voice from a user terminal apparatus, identify an index value from the received uttered voice, and
search a group corresponding to the index value identified from the received uttered voice by comparing the index value of the group with index value identified from the received uttered voice, and to perform the function corresponding to the uttered voice.

11. The voice recognition server according to claim 10 ,
wherein the processor is further configured to:
divide the training text into meaning units;
identify the index value having the meaning units divided by the first text divider as scalar or vector values; and
tag the index value identified by the first meaning index part to each of the meaning units.

12. The voice recognition server according to claim 11 ,
wherein the index value is to be identified using a Skip-Gram method.

13. The voice recognition server according to claim 11 ,
wherein the processor is further configured to:
set a function corresponding to the set group, and to tag the function to the group and store the tagged result.

14. The voice recognition server according to claim 13 ,
wherein the group setting processing circuitry further comprises group index processing circuitry configured to set an index value closest to the average value as the index value of the group.

15. The voice recognition server according to claim 14 ,
wherein the processor is configured to set at least one of: a command processing function, an intention processing function and a sentiment processing function, as the function corresponding to the group, based on a received command.

16. The voice recognition server according to claim 15 ,
wherein the processor is further configured to:
divide a text recognized from the uttered voice into meaning units; and
identify the index value by expressing the meaning units divided from the text as scalar or vector values.

17. The voice recognition server according to claim 16 ,
wherein the processor is configured to select a group corresponding to the identified index value among the pre-stored groups; and to transmit a voice recognition result corresponding to the function of the selected group to the user terminal apparatus.

18. The voice recognition server according to claim 17 ,
wherein the processor, in response to the function of the selected group being the command processing function, is configured to transmit a control command for controlling an operation of the user terminal apparatus to the user terminal apparatus, and
in response to the function of the selected group being the intention processing function or the sentiment processing function, is configure to transmit at least one of: a response message, a control message, and a search result of the uttered voice to the user terminal apparatus.

Assignees

• Samsung Electronics Co Ltd

Inventors

• Lee Kyung-Min
• Kim Il-Hwan
• Park Chi-Youn
• Han Young-Ho
• Kim Nam-Hoon
• Lee Jae-Won

Classifications

• G10L15/1822

  Parsing for meaning understanding · CPC title

• G10L2015/223

  Execution procedure of a spoken command · CPC title

• G10L15/19 Primary

  Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules · CPC title