Apparatus and method for detecting named entity

• US 20070100814A1
• Filed: 08/03/2006
• Published: 05/03/2007
• Est. Priority Date: 10/28/2005
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd. (Samsung Group)
• Original Assignee: Samsung Electronics Co. Ltd. (Samsung Group)
• Inventors: Lee, Jae-won, Kim, Jeong-su, Rim, Hae-chang, Kang, In-ho
• Granted Patent: US 8,655,646 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G10L 15/18 using natural language modelling

Abstract

An apparatus and method for detecting a named-entity. The apparatus includes a candidate-named-entity extraction module that detects a candidate-named-entity based on an initial learning example and feature information regarding morphemes constituting an inputted sentence, the candidate-named-entity extraction module providing a tagged sentence including the detected candidate-named-entity; a storage module that stores information regarding a named-entity dictionary and a rule; and a learning-example-regeneration module for finally determining whether the candidate-named-entity included in the provided sentence is a valid named-entity, based on the named-entity dictionary and the rule, the learning-example-regeneration module providing the sentence as a learning example, based on a determination result, so that a probability of candidate-named-entity detection is gradually updated.

Claims

1. A named-entity detection apparatus comprising:

a candidate-named-entity extraction module detecting a candidate-named-entity based on an initial learning example and feature information regarding morphemes constituting an inputted sentence, the candidate-named-entity extraction module providing a tagged sentence including the detected candidate-named-entity;

a storage module storing information regarding a named-entity dictionary and a rule; and

a learning-example-regeneration module determining whether the candidate-named-entity included in the provided sentence is a valid named-entity, based on the named-entity dictionary and the rule, and providing the sentence as a learning example, based on a determination result, so that a probability of named-entity candidate detection is gradually updated.

2. The named-entity detection apparatus of claim 1, wherein the feature information comprises at least one feature selected from the group consisting of a word feature, a part-of-speech feature, a concept feature, an adjacent verb feature, and an adjacent verb concept feature.

3. The named-entity detection apparatus of claim 1, wherein the candidate-named-entity extraction module determines a class of each morpheme by using a learning model based on the feature information, and provides a tagged sentence by adding determined class information as a tag.

4. The named-entity detection apparatus of claim 3, wherein the learning model comprises a Bayesian learning model.

5. The named-entity detection apparatus of claim 1, wherein the named-entity dictionary comprises unambiguous named-entities.

6. The named-entity detection apparatus of claim 1, wherein the rule comprises an error inspection rule using a meaning string pattern.

7. The named-entity detection apparatus of claim 1, wherein the rule comprises an error inspection rule using an excluded word list.

8. The named-entity detection apparatus of claim 1, wherein the learning example provided by the learning-example-regeneration module is provided as a negative example when the rule is violated.

9. The named-entity detection apparatus of claim 1, wherein the learning-example-regeneration module provides the sentence including the candidate-named-entity as a positive learning example when the candidate-named-entity is included in the named-entity dictionary.

10. A named-entity detection method comprising:

detecting a candidate-named-entity based on an initial learning example and feature information regarding morphemes constituting an inputted sentence and providing a tagged sentence including the detected candidate-named-entity;

finally determining whether the candidate-named-entity included in the provided sentence is a valid named-entity, based on pre-stored named-entity dictionary and rule, and

providing the sentence as a learning example, based on a determination result, so that a probability of candidate-named-entity detection is gradually updated.

11. The named-entity detection method of claim 10, wherein the feature information comprises at least one feature selected from the group consisting of a word feature, a part-of-speech feature, a concept feature, an adjacent verb feature, and an adjacent verb concept feature.

12. The named-entity detection method of claim 10, wherein, in the detecting, a class of each morpheme is determined by using a learning model based on the feature information, and a tagged sentence is provided after adding determined class information as a tag.

13. The named-entity detection method of claim 12, wherein the learning model comprises a Bayesian learning model.

14. The named-entity detection method of claim 10, wherein the named-entity dictionary comprises unambiguous named-entities.

15. The named-entity detection method of claim 10, wherein the rule comprises an error inspection rule using a meaning string pattern.

16. The named-entity detection method of claim 10, wherein the rule comprises an error inspection rule using an excluded word list.

17. The named-entity detection method of claim 10, wherein, in the providing, the learning example is provided as a negative example when the rule is violated.

18. The named-entity detection method of claim 10, wherein, in the providing, the sentence including the candidate-named-entity is provided as a positive learning example when the candidate-named-entity is included in the named-entity dictionary.

23. A computer-readable storage medium encoded with processing instructions for causing a processor to execute the method of claim 10.

19. A named-entity detection system comprising:

a voice-recognition unit converting an utterance into a recognized sentence;

a morpheme analysis unit analyzing morphemes of the recognized sentence; and

a named-entity detection unit detecting named-entities from the analyzed sentence, the named-entity detection unit comprising

a model-learning module learning a model for named-entity extraction from at least one example,

a named-entity candidate extraction module extracting all possible named-entity candidates based on feature information regarding morphemes constituting the recognized sentence and either a model learned by the model-learning module or an initial learning example, and providing a tagged sentence including all the extracted named-entity candidates,

a storage module storing a named-entity dictionary and a rule, and

a learning-example-regeneration module determining whether the at least one named-entity candidate included in the provided tagged sentence is a valid named-entity, based on the named-entity dictionary and the rule, and providing the tagged sentence as a learning example to the model-learning module, based on a determination result, so as to increase a probability of named-entity candidate detection.

20. The system of claim 19, further comprising a sentence-input unit receiving an inputted sentence, wherein the morpheme analysis unit analyzes morphemes of the inputted sentence.

21. A named-entity detection method comprising:

extracting a candidate named-entity, which is a potential named-entity, from a user utterance including a sentence using an initial learning example and feature information regarding morphemes constituting the sentence;

determining whether a candidate named-entity matches an item registered in a dictionary;

verifying the candidate named-entity as a named-entity when the candidate named-entity matches an item registered in the dictionary and providing the candidate named-entity as a new positive learning example to a learning model; and

determining whether the candidate named-entity violates a rule when the candidate named-entity does not match an item registered in the dictionary, providing the candidate named-entity as a new negative learning example to the learning model when the candidate violates a rule, and verifying the candidate named-entity as a named-entity when the candidate named-entity does not violate any rule.

22. The method of claim 21, wherein the rule is an error inspection rule using a meaning string pattern and/or an error inspection rule using an excluded word list.

24. A computer-readable storage medium encoded with processing instructions for causing a processor to execute the method of claim 21.