KR102188564B1 - Method and system for machine translation capable of style transfer - Google Patents

Abstract

A machine translation method and system capable of language conversion are disclosed. The machine translation method according to an embodiment includes receiving text of a target language corresponding to text of a source language from learning data including text of a source language and text of a corresponding target language, the received text of the target language Receiving a token for identifying a language corresponding to the machine translation model, and training the machine translation model to translate the text of the source language into the text of the corresponding target language to which the language identified by the token is applied It may include.

Description

Machine translation method and system capable of language conversion {METHOD AND SYSTEM FOR MACHINE TRANSLATION CAPABLE OF STYLE TRANSFER}

The following description relates to a machine translation method and system capable of language conversion.

Machine translation means using a computer to translate different languages. For example, machine translation may translate text in a first language (eg, English) that is a source language into text in a second language that is a target language (eg Korean). Non-Patent Document 1 relates to a multilingual neural network machine translation system of Google, and discloses a solution capable of translating multiple languages using a single neural network machine translation (NMT). The solution of this non-patent document 1 discloses a technology for specifying a required target language by introducing an artificial token at the beginning of an input sentence without the need to change the model architecture in the standard NMT system. For example, when English text and Korean tokens are input into the machine translation model, the English text is translated into Korean text, and when the English text and Japanese token are input into the machine translation model, the English text may be translated into Japanese text.

On the other hand, one language can have various styles. For example, the basic form of Korean Zombie is largely composed of two types: honorific and semimal. In general, depending on the form of the predicate, honorific and seminal words can be classified, but there are cases where the terms (for example, birthday / birthday (in Korean)) change depending on the honorific and seminal language. . In the prior art, in order to apply the language of the target language, the word language of the already translated text may be changed by applying a rule-based post-processing process to the text translated from the source language to the target language through a machine translation model. However, this prior art has a problem that a separate post-processing converter is required because the machine translation model cannot distinguish the language of the target language.

Non-Patent Document 1: Google's Multilingual Neural Machine Translation System: Enabling Zero-Shot Translation (Melvin Johnson, Mike Schuster,… Nikhil Thorat, 2017.08.21)

It provides a machine translation method and system capable of learning a machine translation model for translating a source language into a target language even to the language of the language.

A machine translation method and system capable of translating a source language into a target language in consideration of the language by using a machine translation model learned up to the language of the language is provided.

Provides a machine translation method and system that can easily tag words for text without changing existing training data for machine translation model learning.

A machine translation method of a computer device including at least one processor, comprising: a target language corresponding to a text of a source language from learning data including text of a source language and text of a corresponding target language by the at least one processor Receiving the text of the input; Receiving, by the at least one processor, a token for identifying a language corresponding to the input text of the target language; And training the machine translation model to translate the text of the source language into the text of the corresponding target language to which the language identified by the token is applied.

According to one aspect, the machine translation method includes, by the at least one processor, providing a translation function for a translation service; Identifying, by the at least one processor, a source language, a target language, and a language selected through the translation function; The selected target language to which the identified language is applied by inputting the text of the selected source language inputted through the translation function and a token identifying the identified language into the learned machine translation model by the at least one processor Generating the text of; And providing, by the at least one processor, the generated text through the translation function.

According to another aspect, the step of checking the source language, the target language, and the language may include managing languages for each target language; Providing, through the translation function, a language selection function for selecting one of the words corresponding to the target language selected through the translation function; And confirming the word selected through the word selection function.

According to another aspect, the machine translation method further includes, by the at least one processor, tagging a token for identifying a language for each target language to the text of the target language included in the learning data, and the token The receiving of the input may include receiving a token tagged to the corresponding target language.

According to another aspect, in the case of Korean, the word font includes a word for honorable words and a word for non-honored words, and the tagging includes a word font of the Korean text included in the learning data using a rule-based classifier. Changing to; And tagging the honorific token for identifying the honorific word to the text whose content has been changed as a result of the change.

According to another aspect, the tagging may further include tagging a non-respectful token for identifying a non-respectful word to the text whose content has not been changed as a result of the change.

In combination with a computer device, there is provided a computer program stored on a computer-readable recording medium for executing the method on the computer device.

A computer-readable recording medium in which a program for executing the method is recorded on a computer device is provided.

A computer device included in a machine translation system, comprising at least one processor embodied to execute an instruction readable by the computer device, by the at least one processor, text in a source language and text in a corresponding target language A text of a target language corresponding to the text of the source language is input from the training data including, and a token for identifying a language corresponding to the text of the input target language is input, and the machine translation model converts the text of the source language. It provides a computer device, characterized in that for training the machine translation model to translate the text of the corresponding target language to which the language identified by the token is applied.

A machine translation model for translating the source language into the target language can be trained to the language of the language.

The source language can be translated into a target language in consideration of the language by using a machine translation model learned even the language of the language.

You can easily tag words for text without changing the existing training data for machine translation model learning.

1 is a diagram showing an example of a network environment according to an embodiment of the present invention.
2 is a block diagram showing an example of a computer device according to an embodiment of the present invention.
3 is a flowchart showing an example of a machine translation method according to an embodiment of the present invention.
4 is a diagram illustrating an example of a process of learning honorific words according to an embodiment of the present invention.
5 is a diagram illustrating an example of a machine translation process according to an embodiment of the present invention.
6 to 8 are diagrams showing examples of a translation function according to an embodiment of the present invention.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings.

The machine translation system according to the embodiments of the present invention may be implemented through a computer device to be described later. At this time, a computer program according to an embodiment of the present invention may be installed and driven in the computer device, and the computer device may perform the machine translation method according to the embodiments of the present invention under control of the driven computer program. . The above-described computer program may be combined with a computer device and stored in a computer-readable recording medium to execute a machine translation method on a computer. According to an embodiment, the machine translation system may be performed through linkage between two or more computer devices.

1 is a diagram showing an example of a network environment according to an embodiment of the present invention. The network environment of FIG. 1 shows an example including a plurality of electronic devices 110, 120, 130, and 140, a plurality of servers 150 and 160, and a network 170. 1 is an example for explaining the present invention, and the number of electronic devices or servers is not limited as in FIG. 1. In addition, the network environment of FIG. 1 is only for describing one example of environments applicable to the embodiments, and the environment applicable to the embodiments is not limited to the network environment of FIG. 1.

The plurality of electronic devices 110, 120, 130, and 140 may be a fixed terminal implemented as a computer device or a mobile terminal. Examples of the plurality of electronic devices 110, 120, 130, 140 include smart phones, mobile phones, navigation, computers, notebook computers, digital broadcasting terminals, personal digital assistants (PDAs), portable multimedia players (PMPs). ), tablet PC, etc. As an example, in FIG. 1, the shape of a smartphone is shown as an example of the electronic device 1 110, but in the embodiments of the present invention, the electronic device 1 110 substantially connects the network 170 using a wireless or wired communication method. Through this, it may mean one of various physical computer devices capable of communicating with other electronic devices 120, 130, and 140 and/or the servers 150 and 160.

The communication method is not limited, and short-range wireless communication between devices as well as a communication method using a communication network (for example, a mobile communication network, a wired Internet, a wireless Internet, a broadcasting network) that the network 170 may include may be included. For example, the network 170 includes a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). , Internet, and the like. In addition, the network 170 may include any one or more of a network topology including a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree or a hierarchical network, etc. Not limited.

Each of the servers 150 and 160 is a computer device or a plurality of computers that communicates with a plurality of electronic devices 110, 120, 130, and 140 through a network 170 to provide commands, codes, files, contents, services, etc. It can be implemented with devices. For example, the server 150 serves as a plurality of electronic devices 110, 120, 130, and 140 connected through the network 170 (for example, translation services, financial services, payment services, social network services, It may be a system that provides a messaging service, a search service, a mail service, a content providing service, etc.).

2 is a block diagram showing an example of a computer device according to an embodiment of the present invention. Each of the electronic devices 110, 120, 130, and 140 described above or each of the servers 150 and 160 may be implemented by the computer apparatus 200 illustrated in FIG. 2.

As illustrated in FIG. 2, the computer device 200 may include a memory 210, a processor 220, a communication interface 230, and an input/output interface 240. The memory 210 is a computer-readable recording medium and may include a permanent mass storage device such as a random access memory (RAM), read only memory (ROM), and a disk drive. Here, a non-destructive large-capacity recording device such as a ROM and a disk drive may be included in the computer device 200 as a separate permanent storage device separated from the memory 210. In addition, an operating system and at least one program code may be stored in the memory 210. These software components may be loaded into the memory 210 from a computer-readable recording medium separate from the memory 210. Such a separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, disk, tape, DVD/CD-ROM drive, and memory card. In another embodiment, software components may be loaded into the memory 210 through a communication interface 230 other than a computer-readable recording medium. For example, software components may be loaded into the memory 210 of the computer device 200 based on a computer program installed by files received through the network 170.

The processor 220 may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input/output operations. Commands may be provided to the processor 220 by the memory 210 or the communication interface 230. For example, the processor 220 may be configured to execute a command received according to a program code stored in a recording device such as the memory 210.

The communication interface 230 may provide a function for the computer device 200 to communicate with other devices (eg, storage devices described above) through the network 170. For example, a request, command, data, file, etc., generated by the processor 220 of the computer device 200 according to a program code stored in a recording device such as the memory 210, is transmitted to the network according to the control of the communication interface 230. 170) can be transferred to other devices. Conversely, signals, commands, data, files, etc. from other devices may be received by the computer device 200 through the communication interface 230 of the computer device 200 via the network 170. Signals, commands, data, etc. received through the communication interface 230 may be transmitted to the processor 220 or the memory 210, and the file, etc. may be a storage medium (described above) that the computer device 200 may further include. Permanent storage).

The input/output interface 240 may be a means for an interface with the input/output device 250. For example, the input device may include a device such as a microphone, a keyboard, or a mouse, and the output device may include a device such as a display or a speaker. As another example, the input/output interface 240 may be a means for interfacing with a device in which input and output functions are integrated into one, such as a touch screen. The input/output device 250 may be configured with the computer device 200 and one device.

Further, in other embodiments, the computer device 200 may include fewer or more components than the components of FIG. 2. However, there is no need to clearly show most of the prior art components. For example, the computer device 200 may be implemented to include at least some of the input/output devices 250 described above, or may further include other components such as a transceiver and a database.

3 is a flowchart showing an example of a machine translation method according to an embodiment of the present invention. The machine translation method according to the present embodiment may be performed by the computer device 200 implementing the server 150 that provides a translation service to the plurality of electronic devices 110, 120, 130, and 140 described above as an example. have. In another embodiment, the machine translation method may be performed by the computer device 200 implementing any one of the plurality of electronic devices 110, 120, 130, and 140. For example, the processor 220 of the computer device 200 may be implemented to execute a code of an operating system included in the memory 210 or a control instruction according to the code of at least one program. Here, the processor 220 may cause the computer device 200 to perform the steps 310 to 380 included in the method of FIG. 3 according to a control command provided by a code stored in the computer device 200. Can be controlled.

In step 310, the computer device 200 may tag a token identifying a language for each target language on the text of the target language included in the training data. The training data may be basically configured to include text of a source language and text of a corresponding target language. For example, for example, if the source language is English and the target language is Korean, the training data may contain the text "Thank you." in Korean, which corresponds to the text "Thank you." in English, in association with each other. I can.

At this time, for example, in the case of Korean, the language may be divided into an honorable language and a non-honorable language. In this case, the computer device 200 may change the word font of the Korean text included in the training data into a non-honored word font using a rule-based classifier. As an example, in a rule-based classifier, rules for respectable and non-honored words are defined in advance as shown in Table 1 below, and the text may be changed according to the predefined rules.

input Polite word Half-word (non-respect) Thank you. Thank you | thank you. | Thank you Thanks | Thank you I I I If If it is If

In this case, the contents of the text that previously included the honorific font may be changed (for example, the text "Thank you" in the honorific font may be changed to the non-honorable text "Thank you"). In other words, texts with changes in their content can be easily distinguished with respectable texts. Accordingly, the computer device 200 may tag the honorific token that identifies the honorific word to the text whose content has been changed as a result of the change. When a token is also used for non-honored words, the computer device 200 may tag a non-honored token for identifying non-honored words to text whose contents are not changed as a result of the change. The token tagging method may vary according to language characteristics. In step 320, the computer device 200 may receive text of a source language and text of a target language corresponding to the text of the source language from the training data. For example, when the source language is English and the target language is Korean, the computer device 200 may receive an English text "Thank you." and a corresponding Korean text "Thank you" from the training data. have.

In step 330, the computer device 200 may receive a token for identifying a word corresponding to the text of the input target language. In this case, as an input token, the computer device 200 may receive a token of a target language tagged to the text of the source language. For example, if the input text is "Thank you", the text of the corresponding Korean may be tagged with a honorific token to identify the honorific word, and the computer device 200 may obtain the honorific token in step 330. Can be input.

In operation 340, the computer device 200 may train the machine translation model so that the machine translation model receives text of the source language and outputs the text of the target language to which the language identified by the token is applied. For example, the machine translation model is trained to translate the text "Thank you." in the source language English into the text "Thank you" in the target language, Korean, while the text "Thank you" in the Korean language through honorific tokens is received. You will be able to learn to have honorific words. In another example, the machine translation model uses the English text "Thank you." together with the Korean text "Thank you." and a non-honored comment token. When translating the English text "Thank you." into a non-honored word, "Thank you." You can learn to translate it as "Thank you." In other words, when the English text "Thank you." and the honorific token as Korean tokens are input as a later learned machine translation model, the machine translation model converts the English text "Thank you." into the honorific Korean text "Thank you." It can be translated as ". In addition, when a non-honored comment token is input as a Korean token along with the English text "Thank you." as the learned machine translation model, the machine translation model converts the English text "Thank you." into the non-honored Korean text "Thank you." Can be translated as "."

In step 350, the computer device 200 may provide a translation function for a translation service. In one embodiment, when the computer device 200 implements the server 150, the translation function is performed by users (for example, a plurality of electronic devices 110, 120, and/or other devices) through an app-based service or a web-based service. 130, 140)). For example, the electronic device 110 may access the server 150 through a computer program installed and driven in the electronic device 110 (eg, a translation app) to receive a translation function. In another embodiment, when the computer device 200 implements any one of a plurality of electronic devices 110, 120, 130, 140, the translation function is a computer program installed and driven in the computer device 200 (example As such, it may be provided to a user of the computer device 200 through a translation app).

In step 360, the computer device 200 may check the source language, target language, and language selected through the translation function. For example, the translation function may include a source language selection function for selecting a source language from a user, a target language selection function for selecting a target language, and a language selection function for selecting a language of the target language. In this case, the computer device 200 may manage words for each target language, and may provide a language selection function for selecting one of the words corresponding to the target language selected through the translation function through the translation function. In this case, the computer device 200 may check the language selected through the language selection function as the language of the target language requested by the user.

In step 370, the computer device 200 inputs the text of the selected source language input through the translation function and the token identifying the identified language into the learned machine translation model to generate the text of the target language to which the identified language is applied. can do. As described above, the machine translation model may be trained to translate the text of the target language into a language identified by the input token when translating the text of the source language into the text of the target language. For example, in translating English text into Korean text, the machine translation model may translate Korean text into honorific words when a honorific token is input, and translate Korean text into non-honesty words when a non-respect token is input.

In step 380, the computer device 200 may provide the generated text through a translation function. For example, when the user selects "English" as the source language, "Korean" as the target language, and "Holy comment language" as the language, the Korean text of the honorific language may be provided as a translation result. In the case of a voice-based translation service, a Korean voice corresponding to a Korean text in an honorable language may be output.

4 is a diagram illustrating an example of a process of learning honorific words according to an embodiment of the present invention. In this embodiment, the machine translation system 400 may include the computer device 200 described above. In this case, the machine translation system 400 may input the text of the source language, the text of the target language, and the tagged token into the machine translation model 420 from the training data 410 to identify the language of the target language. As already described, tokens may be previously tagged for each word of the target language in the learning data 410. Table 2 below shows examples of training data tagged with honorific tokens.

token Text in source language Text in target language <hnr> Your terms are provided for free delivery of all orders worth US $200 or more. Free shipping is provided for orders over US$200 in your payment terms. It's great weather to go picnic. It's a good weather for a picnic. Dear Tony, Dear Tony, <hnr> I will be unable to attend class this Friday. I don't think I will be able to attend class this Friday. As a part of the administrative restructuring As part of management reform … … …

In Table 2, each row may refer to individual learning data, and <hnr> may refer to an honorific token for identifying honorific words. When the machine translation model 420 receives training data including the honorific token, it can be seen that the text of the target language has the honorific word, and can learn this. In Table 2, the text of the target language of the training data that is not tagged with the honorific token is considered as a non-honorable language. On the other hand, the non-honorable token may be directly tagged separately from the honorific token. For example, if there are three or more words, tokens for more than one word need to be tagged. Table 3 below shows examples of learning data tagged with honorific tokens and non-respect tokens.

token Text in source language Text in target language <hnr> Your terms are provided for free delivery of all orders worth US $200 or more. Free shipping is provided for orders over US$200 in your payment terms. <bse> It's great weather to go picnic. It's a good weather for a picnic. <bse> Dear Tony, Dear Tony, <hnr> I will be unable to attend class this Friday. I don't think I will be able to attend class this Friday. <bse> As a part of the administrative restructuring As part of management reform … … …

In Table 3, each row can mean individual learning data, <hnr> means honorific tokens to identify honorific words, and <bse> denotes non-honorific words to identify non-honorific words. Also, for tagging of tokens, the classifier 430 may be utilized. As described through Table 1, the classifier 430 may be used to change the honorable word into the non-honorable word and/or the non-honorable word into the honorable word according to an already set rule. In this case, tagging of the honorific token may change all texts of the target language in the training data into non-honorable words using the classifier 430 as described above. At this time, the classifier 430 can recognize all texts of the target language that are included in the learning data whose contents are changed according to the change of the words, and tag the honorific tokens on the texts of the target language recognized as the honorifics. I can. As an example, Table 4 shows examples of changes in the content of text in the target language.

Text in source language Text in target language Content change Your terms are provided for free delivery of all orders worth US $200 or more. Free shipping is provided for orders over US$200 in your payment terms. ~ Is defined. It's great weather to go picnic. It's a good weather for a picnic. none Dear Tony, Dear Tony, none I will be unable to attend class this Friday. I don't think I will be able to attend class this Friday. I don't think I can. As a part of the administrative restructuring As part of management reform none … … …

Table 4 shows that the contents of the text of the target language in the first row and the text in the target language in the fourth row have changed. In this case, the text of the target language in the first row and the text in the target language in the fourth row may be tagged with honorific tokens. FIG. 5 is a diagram illustrating an example of a machine translation process according to an embodiment of the present invention. to be. The machine translation system 400 may provide a translation function 520 to the user 510. In this case, the user 510 may set a source language, a target language, and a language through the translation function 520 and may input text of the source language. In this case, the text input through the translation function 520 and the tagged token may be transmitted to the machine translation model 420. The machine translation model 420 may translate the input text from the source language into the target language according to the set source language and the target language. In this case, the machine translation model 420 is identified by the token to which the language of the target language is delivered. Translation can be processed to have a word. In addition, the machine translation model 420 may transmit the translated text of the target language to the translation function 520, and the translation result may be transmitted to the user 510 through the translation function 520. Therefore, when the user 510 translates the source language into the target language, the user 510 can receive the translation result of the language of the target language desired by the user 510.

6 to 8 are diagrams showing examples of a translation function according to an embodiment of the present invention.

6 shows an example of a translation function screen 600 provided based on an app or web. For example, the translation function screen 600 may be a terminal screen of a user who uses a translation service. In this case, the first dotted box 610 shows an example of a user interface for selecting a source language. When an area of the first dotted box 610 is selected (for example, when a corresponding area is touched in a touch screen environment or a mouse click occurs in a corresponding area in a PC environment), a list of selectable languages may be displayed, The user can select the desired source language from the list. Also, the source language input area 620 may be a user interface for inputting text of a source language selected by the user. The translation function screen 600 shows an example in which the user inputs the text "Thank you." of the Suso language "English". According to an embodiment, when a user inputs text, a source language corresponding to the input text may be automatically selected. The second dotted box 630 shows an example of a user interface for selecting a target language, and the third dotted box 640 shows an example of a user interface for selecting a word of the selected target language. On the translation function screen 600, as the user selects "Korean" as the target language and selects the area of the third dotted box 640, a list 650 of words such as "honored words" and "half words" is displayed. The displayed example is shown. The user can select a desired word from the list 650. Further, on the translation function screen 600, a target language output area 660 for displaying a translation result is shown. When the user selects the translate button 670, translation may proceed according to the setting. Depending on the embodiment, the translation function includes a function for listening to a pronunciation corresponding to an input and/or output text, a function for listening to a pronunciation corresponding to an output text, a function for copying input and/or output text, text or Various additional functions, such as a function for sharing translation results, may be further included.

In FIG. 7, as the user selects the translate button 670 while selecting the "honored word" as shown in the fourth dotted box 710, as a result of translation of the English text "Thank you." The Korean text "Thank you" to which is applied is shown in an example in which the target language output area 660 is displayed.

On the other hand, in FIG. 8, as the user selects the translate button 670 while selecting the "banmal" as shown in the fifth dotted box 810, the translation result for the English text "Thank you." An example is shown in which the Korean text "Thank you." to which a semi-mal (non-honorable) language is applied is displayed in the target language output area 660.

In the above embodiments, only honorable and semimal (non-honored words) are described as examples of Korean language, but in detail there are various words such as "Hao", "Hao", and "Haeara", such as tagging and machine. These detailed words may be used to train the translation model, and these words may differ depending on the target language.

In addition, although the above embodiments describe a text-based translation service, even when such text is converted into speech to provide a real-time interpretation service, the text-to-speech is basically after the input voice and before the output voice. It will be easily understood by those of ordinary skill in the art that since the transformation is utilized, the language of the language translated in the text section can be considered.

As described above, according to embodiments of the present invention, a machine translation model for translating a source language into a target language can be trained to the language of the language. In addition, it is possible to translate the source language into the target language in consideration of the language by using a machine translation model learned even the language of the language. In addition, it is possible to easily tag words for text without changing the existing training data for machine translation model learning.

The system or device described above may be implemented as a hardware component or a combination of a hardware component and a software component. For example, the devices and components described in the embodiments are, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA). , A programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions, such as one or more general purpose computers or special purpose computers. The processing device may execute an operating system (OS) and one or more software applications executed on the operating system. In addition, the processing device may access, store, manipulate, process, and generate data in response to the execution of software. For the convenience of understanding, although it is sometimes described that one processing device is used, one of ordinary skill in the art, the processing device is a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it may include. For example, the processing device may include a plurality of processors or one processor and one controller. In addition, other processing configurations are possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of these, configuring the processing unit to behave as desired or processed independently or collectively. You can command the device. Software and/or data may be interpreted by a processing device or to provide instructions or data to a processing device, of any type of machine, component, physical device, virtual equipment, computer storage medium or device. Can be embodyed in The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like alone or in combination. The medium may be one that continuously stores a program executable by a computer, or temporarily stores a program for execution or download. In addition, the medium may be a variety of recording means or storage means in a form in which a single piece of hardware or several pieces of hardware are combined. The medium is not limited to a medium directly connected to a computer system, but may be distributed on a network. Examples of media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magnetic-optical media such as floptical disks, and And a ROM, RAM, flash memory, and the like, and may be configured to store program instructions. In addition, examples of other media include an app store that distributes applications, a site that supplies or distributes various software, and a recording medium or storage medium managed by a server. Examples of program instructions include not only machine language codes such as those produced by a compiler but also high-level language codes that can be executed by a computer using an interpreter or the like.

As described above, although the embodiments have been described by the limited embodiments and drawings, various modifications and variations are possible from the above description by those of ordinary skill in the art. For example, the described techniques are performed in a different order from the described method, and/or components such as a system, structure, device, circuit, etc. described are combined or combined in a form different from the described method, or other components Alternatively, even if substituted or substituted by an equivalent, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and equivalents to the claims fall within the scope of the following claims.

Claims (12)

In the machine translation method of a computer device comprising at least one processor,
Receiving, by the at least one processor, the text of the source language and the text of the target language corresponding to the text of the source language from the training data including the text of the source language and the text of the corresponding target language for learning the machine translation model;
Receiving, by the at least one processor, a token for identifying a language corresponding to the received text of the target language for learning the machine translation model; And
Training the machine translation model so that the machine translation model translates the text of the source language into the text of the corresponding target language to which the language identified by the token is applied.
Machine translation method comprising a. The method of claim 1,
Providing, by the at least one processor, a translation function for a translation service;
Identifying, by the at least one processor, a source language, a target language, and a language selected through the translation function;
The selected target language to which the identified language is applied by inputting the text of the selected source language inputted through the translation function and a token identifying the identified language into the learned machine translation model by the at least one processor Generating the text of; And
Providing the generated text through the translation function, by the at least one processor
Machine translation method further comprising a. The method of claim 2,
The step of confirming the source language, target language and language,
Managing words for each target language;
Providing, through the translation function, a language selection function for selecting one of the words corresponding to the target language selected through the translation function; And
Checking the font selected through the font selection function
Machine translation method comprising a. The method of claim 1,
Tagging, by the at least one processor, a token identifying a language for each target language in the text of the target language included in the learning data
Including more,
The step of receiving the token,
Machine translation method, characterized in that receiving input of the token tagged to the corresponding target language. The method of claim 4,
In the case of Korean, the words include honorable and non-honored words,
The tagging step,
Changing the word font of the Korean text included in the learning data into a non-honored word font using a rule-based classifier; And
Tagging the honorific token for identifying the honorific word to the text whose content is changed as a result of the change
Machine translation method comprising a. The method of claim 5,
The tagging step,
Tagging a non-honorable token identifying a non-honorable word to text whose content has not been changed as a result of the change
Machine translation method further comprising a. A computer program combined with a computer device and stored in a computer-readable recording medium for executing the method of claim 1 on a computer device. A computer-readable recording medium storing a computer program for executing the method of claim 1 on a computer device. In the computer device included in the machine translation system,
At least one processor embodied to execute instructions readable by the computer device
Including,
By the at least one processor,
The text of the source language and the text of the target language corresponding to the text of the source language are input for learning the machine translation model from training data including the text of the source language and the text of the corresponding target language,
By the at least one processor, a token for identifying a language corresponding to the input text of the target language is input for learning the machine translation model,
Training the machine translation model so that the machine translation model translates the text of the source language into the text of the corresponding target language to which the language identified by the token has been applied.
Computer device, characterized in that. The method of claim 9,
By the at least one processor,
Provide translation functions for translation services,
Check the source language, target language and language selected through the translation function,
Inputting the text of the selected source language input through the translation function and a token identifying the identified language into the learned machine translation model to generate the text of the selected target language to which the identified language is applied,
Providing the generated text through the translation function
Computer device, characterized in that. The method of claim 10,
By the at least one processor,
Manage the languages of each target language,
Providing a language selection function for selecting one of the languages corresponding to the target language selected through the translation function through the translation function,
Checking the font selected through the font selection function
Computer device, characterized in that. The method of claim 9,
By the at least one processor,
Tagging a token identifying a language for each target language to the text of the target language included in the learning data,
To receive the token input, receiving a token tagged to the text of the target language
Computer device, characterized in that.

Images