KR-20260065788-A - System, Apparatus and Method For Processing Natural Language, and Computer Readable Recording Medium

Abstract

The present invention relates to a natural language processing system, a natural language processing device, a natural language processing method, and a computer-readable recording medium. A natural language processing system according to an embodiment of the present invention may include a user device that receives a sentence of compound or complex sentences, and a natural language processing device that generates a plurality of control commands for the user device based on whether an action intended by the user can be performed sequentially in the sentence of compound or complex sentences received from the user device.

Inventors

• 김경덕
• 정지혜

Assignees

• 삼성전자주식회사

Dates

Publication Date

20260511

Application Date

20260428

Claims (20)

1. In electronic devices, Communications Department; Input receiving unit; and Includes a processor; The above processor is, A user voice input containing a first sentence of a first language and a second sentence of a second language sequentially is received through the input receiving unit, and Based on the user voice input transmitted to the server through the communication unit, text corresponding to the user voice input is obtained, and A first function corresponding to the first sentence above and a second function corresponding to the second sentence above are executed sequentially, and Providing first information regarding the execution result of the first function in the first language, and An electronic device that provides second information regarding the result of executing the second function in the second language.
2. In paragraph 1, The above user voice input is, An electronic device comprising a conjunction connecting the first sentence and the second sentence.
3. In paragraph 1, The above user voice input is, An electronic device comprising the first sentence and the second sentence input consecutively at regular time intervals without including conjunctions.
4. In paragraph 1, The above processor is, An electronic device that performs only the first function if the performance of the second function is impossible due to the performance of the first function.
5. In paragraph 1, The above processor is, An electronic device that provides first guide information corresponding to the impossibility of performing the second function when the performance of the first function is impossible due to the performance of the second function.
6. In paragraph 1, The above processor is, If the performance of the second function is impossible due to the performance of the first function, output second guide information corresponding to the inquiry to perform only the first function, and An electronic device that performs only the first function when user voice input is received based on the second guide information above.
7. In paragraph 1, The above processor is, An electronic device that receives user input including the first sentence and the second sentence in text form instead of the user voice input.
8. In paragraph 1, The above processor is, An electronic device that provides the first information and the second information in the same language, instead of providing the first information in the first language and providing the second information in the second language.
9. In paragraph 1, The above processor is, An electronic device that provides third guide information corresponding to said additional user input when additional user input is required to perform one of the first function and the second function.
10. In Paragraph 9, The above third guide information is, An electronic device comprising information for selecting one of a plurality of search results related to the performance of the above-mentioned function.
11. In a method for controlling an electronic device, A step of receiving user voice input that sequentially includes a first sentence of a first language and a second sentence of a second language; A step of obtaining text corresponding to the user voice input based on the user voice input transmitted to the server; A step of sequentially executing a first function corresponding to the first sentence and a second function corresponding to the second sentence; A step of providing first information regarding the result of executing the first function in the first language; and A control method comprising the step of providing second information regarding the execution result of the second function in the second language.
12. In Paragraph 11, The above user voice input is, A control method comprising a conjunction connecting the first sentence and the second sentence.
13. In Paragraph 11, The above user voice input is, A control method comprising the first sentence and the second sentence input consecutively at regular time intervals without including conjunctions.
14. In Paragraph 11, The above control method is, A control method comprising: a step of performing only the first function if the performance of the second function is impossible due to the performance of the first function.
15. In Paragraph 11, The above control method is, A control method comprising: a step of providing first guide information corresponding to the impossibility of performing the second function if the performance of the first function is impossible due to the performance of the second function.
16. In Paragraph 11, The above control method is, If the performance of the second function is impossible due to the performance of the first function, a step of outputting second guide information corresponding to the inquiry to perform only the first function of the imagination device; and A control method comprising: a step of performing only the first function when user voice input is received based on the second guide information above.
17. In Paragraph 11, The above control method is, A control method comprising the step of receiving user input including the first sentence and the second sentence in text form instead of the user voice input.
18. In Paragraph 11, The above control method is, A control method comprising the step of providing the first information in the first language and the second information in the second language, instead of providing the first information and the second information in the second language.
19. In Paragraph 11, The above control method is, A control method comprising the step of providing third guide information corresponding to the additional user input when additional user input is required to perform one of the first function and the second function.
20. In Paragraph 19, The above third guide information is, A control method comprising information for selecting one of a plurality of search results related to the performance of the above-mentioned function.

Description

System, Apparatus and Method For Processing Natural Language, and Computer Readable Recording Medium The present invention relates to a natural language processing system, a natural language processing device, a natural language processing method, and a computer-readable recording medium. More specifically, it relates to a natural language processing system, a natural language processing device, a natural language processing method, and a computer-readable recording medium that determine appropriate system actions for user utterances in the form of complex language, such as voice conversation systems, question-and-answer systems, and chat systems, and selectively process system actions based on whether multiple actions intended by the user can be processed sequentially in a system, such as a video display device. Generally, machine translation refers to the process by which a computer system automatically converts a natural language sentence F of an input language into a natural language sentence E of a target language. Among machine translation techniques, statistical machine translation trains a machine translation model based on training data and performs machine translation based on the trained model. More specifically, it is the process of finding the value E that maximizes the probability Pr(E|F) of E given F. In other words, E is the best translation result for F. This can be expressed as in <Equation 1>. By applying Bayes' rule to the above <Equation 1> and decomposing Pr(E｜F), <Equation 2> can be obtained. Here, Pr(F｜E) represents the probability that a given E is translated into F by the translation model, indicating how appropriate it is to translate E into F. The translation model is trained based on training data for bilingual languages. Here, Pr(E) represents the probability that E appears in the language model and indicates how natural E is. The language model is trained based on training data for a single language (monolingual). Conventional natural language processing systems analyze morphological information, syntactic structure, and semantics from input sentences. In this case, a single input sentence is either a basic sentence of the smallest size or a sentence composed of multiple basic sentences, i.e., a compound sentence. The basic sentences that make up a compound sentence are connected to each other in various forms. For example, let's assume a natural language processing system that recognizes and executes voice commands related to TV programs. A TV user can utter the compound sentence "Record OCN news and show me Family Guy" using a natural language processing system, in which the basic sentences "Record OCN news" and "Show me Family Guy" are connected by the conjunction "and". In addition, in some languages such as Korean, sentences are modified when connected by conjunctions. A TV user can utter a complex sentence to a natural language processing system, "Record Infinite Challenge and play 1 Night 2 Days," in which the basic sentences "Record Infinite Challenge" and "Play 1 Night 2 Days" are connected by the conjunction "go." Meanwhile, a TV user can utter a compound sentence, "Record OCN news show me Family Guy," using a natural language processing system; this compound sentence occurs when the TV user utters two sentences consecutively without a conjunction. However, conventional natural language processing systems have difficulty processing such complex sentences, which leads to a problem of degraded performance. The purpose of the embodiments of the present invention is to provide a natural language processing system, a natural language processing device, a natural language processing method, and a computer-readable recording medium, wherein, in a natural language processing system such as a voice conversation system, a question-and-answer system, and a chat system, appropriate system actions are determined for user utterances in the form of compound words, and system actions are selectively processed according to whether multiple actions intended by the user can be processed sequentially in a system such as a video display device. A natural language processing system according to an embodiment of the present invention includes a user device that receives a sentence of a compound or complex sentence, and a natural language processing device that generates a plurality of control commands for the user device based on whether an action intended by the user can be performed sequentially in the compound or complex sentence received from the user device. Additionally, a natural language processing device according to an embodiment of the present invention includes a storage unit that stores operation information related to whether a plurality of actions intended by a user can be performed sequentially in a compound or complex sentence, a communication interface unit that receives a compound or complex sentence spoken by a user, and a natural language processing unit that determine