Artificial intelligence based method and apparatus for classifying voice-recognized text

• US 10,762,901 B2
• Filed: 08/03/2018
• Issued: 09/01/2020
• Est. Priority Date: 08/23/2017
• Status: Active Grant

• Current Assignee: Beijing Baidu Netcom Science and Technology Company Limited (Baidu Incorporated)
• Original Assignee: Beijing Baidu Netcom Science and Technology Company Limited (Baidu Incorporated)
• Inventors: Liang, Yichuan, Ling, Guang, Lin, Yingzhan, Wang, Wenmin, Xie, Zeying, Zhang, Yin, Xu, Wei, Zhou, Chao
• Primary Examiner(s): Chawan, Vijay B
• Application Number:
  US16/054,970
• Publication Number:
  US 20190066675A1
• Time in Patent Office: 760 Days
• Field of Search: 707999006, 707740, 707749, 707768, 707771, 707 0003, 707999005, 707999104, 715234, 715205, 704 9, 704240, 704257, 704231, 704235, 704260, 706 10, 706 12, 706 20, 706 14, 706 45, 706 46
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06F 3/167 Audio in a user interface, e.g. using voice commands for navigating, audio feedback
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/193 Formal grammars, e.g. finite state automata, context free grammars or word networks
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/0633 using lexical or orthographic knowledge sources
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: G10L 2015/228 of application context

Abstract

Embodiments of the present disclosure disclose an artificial intelligence based method and apparatus for classifying a voice-recognized text. A specific embodiment of the method includes: acquiring a current interactive text of a voice query from a user; analyzing the current interactive text using a lexical analyzer to obtain a current lexical structure; determining whether the current lexical structure matches a template of a category in a classifier; and classifying, if the current lexical structure matches the template of the category in the classifier, the current interactive text corresponding to the current lexical structure into the category belonging to the matched template. The embodiment can fast classify texts, effectively reduce the magnitude of manually annotated texts, and improve the annotation efficiency in intelligent voice interaction services.

Claims

1. An artificial intelligence based method for classifying a voice-recognized text, the method comprising:

acquiring a current interactive text of a voice query from a user;

analyzing the current interactive text using a lexical analyzer to obtain a current lexical structure;

determining whether the current lexical structure matches a template of a given category in a classifier; and

classifying, if the current lexical structure matches the template of the given category in the classifier, the current interactive text corresponding to the current lexical structure into the given category belonging to the matched template, wherein the template in the given category includes a plurality of keywords, and each of the keywords of the template in the given category is determined by;

acquiring the given category for a predetermined interaction scenario;

acquiring a plurality of pieces of to-be-generalized interactive text having same or similar meanings in the given category;

generating one generalized keyword matching the pieces of to-be-generalized interactive text having same or similar meanings in the given category; and

using the generalized keyword as the keyword of the template in the given category.

2. The method according to claim 1, wherein

the to-be-generalized interactive text has a frequency of occurrence higher than a predetermined frequency in the category.

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

classifying, if the current lexical structure fails to match the template of the category in the classifier, the current interactive text corresponding to the current lexical structure into a to-be-classified category.

4. The method according to claim 1, further comprising:

acquiring a to-be-processed interactive text having a frequency of occurrence higher than a preset frequency in the to-be-classified category;

receiving a manually classifying operation of the to-be-processed interactive text; and

classifying the to-be-processed interactive text into a category corresponding to the classifying operation in the classifier.

5. The method according to claim 4, wherein the classifying the to-be-processed interactive text into a category corresponding to the classifying operation in the classifier comprises:

classifying, if there is the category corresponding to the classifying operation in the classifier, the to-be-processed interactive text directly into the category corresponding to the classifying operation in the classifier.

6. The method according to claim 4, wherein the classifying the to-be-processed interactive text into a category corresponding to the classifying operation in the classifier comprises:

establishing, if there is no category corresponding to the classifying operation in the classifier, the category corresponding to the classifying operation in the classifier; and

classifying the to-be-processed interactive text into the category corresponding to the classifying operation in the classifier.

7. The method according to claim 4, further comprising:

receiving manual generalization of the to-be-processed interactive text classified into the category in the classifier; and

updating the template of the category in the classifier based on a generalization result.

8. The method according to claim 1, wherein the current lexical structure including a plurality of pieces matching all keywords of the template is classified into the given category.

9. An artificial intelligence based apparatus for classifying a voice-recognized text, the apparatus comprising:

at least one processor; and

a memory storing instructions, the instructions when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising;

acquiring a current interactive text of a voice query from a user;

analyzing the current interactive text using a lexical analyzer to obtain a current lexical structure;

determining whether the current lexical structure matches a template of a given category in a classifier; and

classifying, if the current lexical structure matches the template of the given category in the classifier, the current interactive text corresponding to the current lexical structure into the given category belonging to the matched template, wherein the template in the given category includes a plurality of keywords, and each of the keywords of the template in the given category is determined by;

acquiring the given category for a predetermined interaction scenario;

acquiring a plurality of pieces of to-be-generalized interactive text having same or similar meanings in the given category;

generating one generalized keyword matching the pieces of to-be-generalized interactive text having same or similar meanings in the given category; and

using the generalized keyword as the keyword of the template in the given category.

10. The apparatus according to claim 9, wherein

the to-be-generalized interactive text has a frequency of occurrence higher than a predetermined frequency in the category.

11. The apparatus according to claim 9, wherein the operations further comprise:

classifying, if the current lexical structure fails to match the template of the category in the classifier, the current interactive text corresponding to the current lexical structure into a to-be-classified category.

12. The apparatus according to claim 9, wherein the operations further comprise:

acquiring a to-be-processed interactive text having a frequency of occurrence higher than a preset frequency in the to-be-classified category;

receiving a manually classifying operation of the to-be-processed interactive text; and

classifying the to-be-processed interactive text into a category corresponding to the classifying operation in the classifier.

13. The apparatus according to claim 12, wherein the classifying the to-be-processed interactive text into a category corresponding to the classifying operation in the classifier comprises:

classifying, if there is the category corresponding to the classifying operation in the classifier, the to-be-processed interactive text directly into the category corresponding to the classifying operation in the classifier.

14. The apparatus according to claim 12, wherein the classifying the to-be-processed interactive text into a category corresponding to the classifying operation in the classifier comprises:

establishing, if there is no category corresponding to the classifying operation in the classifier, the category corresponding to the classifying operation in the classifier; and

classifying the to-be-processed interactive text into the category corresponding to the classifying operation in the classifier.

15. The apparatus according to claim 12, wherein the operations further comprise:

receiving manual generalization of the to-be-processed interactive text classified into the category in the classifier; and

updating the template of the category in the classifier based on a generalization result.

16. A non-transitory computer readable storage medium storing a computer program, wherein the computer program, when executed by a processor, cause the processor to perform operations, the operation comprising:

acquiring a current interactive text of a voice query from a user;

analyzing the current interactive text using a lexical analyzer to obtain a current lexical structure;

determining whether the current lexical structure matches a template of a given category in a classifier; and

classifying, if the current lexical structure matches the template of the given category in the classifier, the current interactive text corresponding to the current lexical structure into the given category belonging to the matched template, wherein the template in the given category includes a plurality of keywords, and each of the keywords of the template in the given category is determined by;

acquiring the given category for a predetermined interaction scenario;

acquiring a plurality of pieces of to-be-generalized interactive text having same or similar meanings in the given category;

generating one generalized keyword matching the pieces of to-be-generalized interactive text having same or similar meanings in the given category; and

using the generalized keyword as the keyword of the template in the given category.