Speech processing method and apparatus therefor

1. A speech processing method, comprising:

collecting a user's spoken utterance including a query;

generating a query text as a text conversion result for the user's spoken utterance including the query;

searching whether there is a query text-spoken response utterance set including a spoken response utterance for the query text in a database which is constructed in advance; and

when there is a query text-spoken response utterance set including a spoken response utterance for the query text in the database, providing the spoken response utterance included in the query text-spoken response utterance set,

wherein the searching of whether there is a query text-spoken response utterance set includes:

searching a query text-spoken response utterance set group in which spoken response utterances responsive to the query text are clustered in the database; and

randomly determining a query text-spoken response utterance set in the query text-spoken response utterance set group to be provided.

2. The speech processing method according to claim 1 , wherein the generating of a query text includes:

transmitting the user's spoken utterance including the query to an external server; and

receiving, from the external server, a conversion result corresponding to the query text for the user's spoken utterance including the query.

3. The speech processing method according to claim 1 , further comprising:

after the providing of the spoken response utterance, updating information about a query text-spoken response utterance set in the database.

4. The speech processing method according to claim 3 , wherein the updating in the database includes:

updating the information about a query text-spoken response utterance set in the database or deleting the information about a query text-spoken response utterance set from the database, based on at least one of a relative providing frequency of a spoken response utterance included in a query text-spoken response utterance set, an available storage capacity of the database, or a predetermined number of updated query text-spoken response utterance sets.

5. The speech processing method according to claim 1 , further comprising:

after the providing of the spoken response utterance,

analyzing a relative providing frequency history of query text-spoken response utterance sets stored in the database for every predetermined period of time; and

deleting, from the database, a query text-spoken response utterance set having a relatively low relative providing frequency among the query text-spoken response utterance sets, as an analysis result.

6. The speech processing method according to claim 1 , further comprising:

when there is no query text-spoken response utterance set including a spoken response utterance for the query text in the database, analyzing an utterance intention of the query text;

generating a new response text for the analyzed utterance intention of the query text;

generating a new spoken response utterance as a speech conversion result for the new response text; and

providing the new spoken response utterance.

7. The speech processing method according to claim 6 , wherein the analyzing of an utterance intention of the query text includes:

analyzing the utterance intention of the query text by performing syntactic analysis or semantic analysis on the query text.

8. A speech processing apparatus, comprising:

at least one processor configured to:

collect a user's spoken utterance including a query,

generate a query text as a text conversion result for the user's spoken utterance including the query,

search whether there is a query text-spoken response utterance set including a spoken response utterance for the query text in a database which is constructed in advance, and

when there is a query text-spoken response utterance set including a spoken response utterance for the query text in the database, provide the spoken response utterance included in the query text-spoken response utterance set,

wherein the processor is further configured to search a query text-spoken response utterance set group in which spoken response utterances responsive to the query text are clustered in the database, and randomly determine a query text-spoken response utterance set in the query text-spoken response utterance set group to be provided.

9. The speech processing apparatus according to claim 8 , wherein the at least one processor is further configured to transmit the user's spoken utterance including the query to an external server, and receive, from the external server, a conversion result corresponding to the query text for the user's spoken utterance including the query.

10. The speech processing apparatus according to claim 8 , wherein the at least one processor is further configured to:

update information about a query text-spoken response utterance set in the database, after the providing of the spoken response utterance.

11. The speech processing apparatus according to claim 10 , wherein the at least one processor is further configured to update the information about a query text-spoken response utterance set in the database or delete the information about a query text-spoken response utterance set from the database, based on at least one of a relative providing frequency of a spoken response utterance included in a query text-spoken response utterance set, an available storage capacity of the database, or a predetermined number of updated query text-spoken response utterance sets.

12. The speech processing apparatus according to claim 8 , wherein the at least one processor is further configured to, after the providing of the spoken response utterance, analyze a relative providing frequency history of query text-spoken response utterance sets stored in the database for every predetermined period of time, and delete, from the database, a query text-spoken response utterance set having a relatively low relative providing frequency among the query text-spoken response utterance sets, as an analysis result.

13. The speech processing apparatus according to claim 8 , wherein the at least one processor is further configured to:

when there is no query text-spoken response utterance set including a spoken response utterance for the query text in the database, analyze an utterance intention of the query text, generate a new response text for the analyzed utterance intention of the query text, generate a new spoken response utterance as a speech conversion result for the new response text, and provide the new spoken response utterance.

14. The speech processing apparatus according to claim 13 , wherein the at least one processor is further configured to analyze the utterance intention of the query text by performing syntactic analysis or semantic analysis on the query text.

15. A speech processing apparatus, comprising:

at least one processor configured to:

collect a user's spoken utterance including a query,

generate a query text as a text conversion result for the user's spoken utterance including the query,

search a query text-spoken response utterance set group in which spoken response utterances responsive to the query text are clustered in a database, and randomly determine a query text-spoken response utterance set in the query text-spoken response utterance set group to be provided, and

provide a spoken response utterance included in the randomly determined query text-spoken response utterance set.