KR-20260066670-A - Interpretation Method and Terminal Device Having Domain-Specific Functions Based on Stepwise Keyword Extraction and Ontology

Abstract

The present invention relates to a context-based specialized terminology optimization interpretation method and a terminal device implemented in a communication terminal. The interpretation method according to the present invention comprises: (a) converting a speaker's voice input into text data; (b) extracting a first keyword (K1) associated with a specialized field within the text data; (c) extracting a second keyword (K2) having a semantic co-occurrence relationship or an ontology-based hierarchical relationship with the first keyword (K1); (d) specifying a target specialized field (D) among a plurality of specialized field categories based on a combination of the first and second keywords; (e) calling a specialized terminology dictionary mapped to the specified target specialized field (D) to select a translation term; and (f) generating and outputting a translation sentence reflecting the selected specialized term. According to the present invention, compared to a single keyword method, errors in mistranslation of polysemous words are drastically reduced, translation terms exclusively optimized for the specialized field are provided, and a high-reliability specialized interpretation service capable of adaptive learning through real-time registration of unknown terms can be provided.

Inventors

• 안범주

Assignees

• 안범주

Dates

Publication Date

20260512

Application Date

20260330

Claims (1)

1. As a context-based technical term optimization interpretation method implemented in a communication terminal, (a) A step of receiving voice input from a speaker and converting it into text data; (b) a step of extracting a first keyword associated with a specialized field from the converted text data above; (c) After extracting the first keyword, a step of extracting a second keyword from the text data that has an association with the first keyword; (d) a step of specifying a target professional field of conversation among a plurality of professional field categories based on a combination of the first keyword and the second keyword; (e) a step of selecting terms within a sentence to be translated by calling a specialized term dictionary pre-mapped to the aforementioned specific target specialized field; and (f) a step of generating and outputting a sentence translated into the recipient's language by reflecting the selected technical terms above; An interpretation method characterized by including

Description

Interpretation Method and Terminal Device Having Domain-Specific Functions Based on Stepwise Keyword Extraction and Ontology The present invention relates to a real-time interpretation method and a terminal device implemented in a communication terminal, and more specifically, to a context-based specialized terminology optimization interpretation method and a terminal device for performing the same, wherein a speaker's voice input is converted into text, a first keyword and a second keyword are extracted stepwise, and a target specialized field of conversation is precisely specified based on a combination of the extracted keywords, thereby selectively applying a terminology dictionary optimized for the relevant specialized field to generate and output a translated sentence. As the global business environment becomes more sophisticated, the demand for real-time interpretation in specific specialized fields such as medicine, law, engineering, finance, and semiconductors is rapidly increasing. While conventional general-purpose Neural Machine Translation (NMT)-based interpretation systems demonstrate significant performance in translating everyday language, they have a fundamental limitation that causes critical translation errors in advanced conversational situations where specialized terminology appears frequently. A core challenge in specialized translation is the problem of polysemy, where a single word or expression carries different meanings across multiple fields. For instance, the English word "Cell" must be translated as "cell" in biology, "battery" in energy engineering, and "solitary confinement" in prisons. Conventional systems often fail to precisely analyze the context of sentences containing such polysemy, frequently resorting to selecting the most common translation based on statistical frequency or making random choices. This can lead to serious consequences, such as misdiagnosis in medical settings, legal disputes, and misunderstandings in technical negotiations. Furthermore, conventional domain-specific translation systems adopt methods in which the user manually selects the translation field in advance or determines the field by relying on single keyword frequency analysis. The former method fails to flexibly handle field switching when multiple specialized fields are mixed within a single conversation session, while the latter method suffers from a problem where the accuracy of field determination is significantly reduced when a single keyword alone cannot sufficiently specify the field—that is, when the keyword is a general term used commonly across multiple fields. Therefore, there is a need for a technology that can dramatically improve the precision of interpretation by first extracting a first keyword related to the field of expertise from voice input, then additionally extracting a second keyword based on semantic and hierarchical associations with the first keyword, and then definitively identifying the target field of expertise through a combination of the two keywords to exclusively select terms optimized for the field of expertise. FIG. 1 is an internal configuration block diagram of a specialized field optimization interpretation terminal device (100) according to one embodiment of the present invention. FIG. 2 is an overall processing flowchart of a context-based specialized terminology optimization interpretation method according to an embodiment of the present invention, illustrating the sequential execution process of steps (a) to (f). Figure 3 is an example of a second keyword (K2) determination logic based on semantic co-occurrence analysis and ontology, illustrating a branching structure by field based on the first keyword (K1), 'Cell'. Figure 4 is a conceptual diagram of a target expertise determination mechanism based on the weighted sum of domain relevance score (DRS) and contextual weight (CW). FIG. 5 is an example of a user interface (UI) that presents candidate fields to a user through the output unit (160) of a communication terminal (100) when the reliability threshold (Th) is not met. Preferred embodiments of the present invention will be described in detail below with reference to the attached drawings. Prior to this, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings. Instead, based on the principle that the inventor can appropriately define the concepts of terms to best describe their invention, they should be interpreted in a meaning and concept consistent with the technical spirit of the present invention. Therefore, the embodiments described in this specification and the configurations illustrated in the drawings are merely some of the most preferred embodiments of the present invention and do not represent all of the technical spirit of the present invention. It should be understood that various equivalents and modifications capable of replacing them may exist at the tim