Translation device, translation system, and translation method

Abstract

To provide a translation device 110 capable of displaying an easily understandable translation even when an original text contains words that are difficult to translate literally. A translation device 110 according to the present disclosure includes an original text acquisition unit 10 that acquires an original text to be translated; an annotation target extraction unit 20 that inputs the original text acquired by the original text acquisition unit 10 to an artificial intelligence unit 200 and acquires and outputs annotation target phrases contained in the original text that are difficult to translate from the artificial intelligence unit 200; an annotation addition unit 30 that acquires annotation information corresponding to the annotation target phrases output by the annotation target extraction unit 20 and generates an annotated original text by adding the annotation information to the annotated target phrases; a translation unit 40 that translates the annotated original text generated by the annotation addition unit 30 into a language to be translated, thereby generating an annotated translation; and a display processing unit 50 that displays the annotated translation translated by the translation unit 40 on a display device 400.

Description

[Technical Field]

[0001] The present disclosure relates to a translation device, a translation system, and a translation method. [Background technology]

[0002] There is a translation device that translates an original text to be translated. As an example of such a translation device, Patent Document 1 discloses a translation device as a conference support tool that displays a screen containing an original text, which is a sentence entered in one of a plurality of languages ​​set as languages ​​to be used in the conference, and a translation of the original text into the remaining languages ​​of the plurality of languages. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-190052 Summary of the Invention [Problem to be solved by the invention]

[0004] The device in Patent Document 1 has a problem in that it cannot display an easily understandable translation when the original text contains words that are difficult to translate literally. For example, when the original text contains uncommon words such as company jargon, even if the original text is translated literally, there is a possibility that the meaning will not be conveyed correctly.

[0005] The present disclosure has been made to solve the above-mentioned problems, and aims to display a translation that is easily understandable even when the original text contains words that are difficult to translate literally. [Means for solving the problem]

[0006] The translation device according to the present disclosure includes an original text acquisition unit that acquires an original text to be translated, an annotation target extraction unit that inputs the original text acquired by the original text acquisition unit into an artificial intelligence unit and acquires and outputs annotation target phrases contained in the original text that are difficult to translate from the artificial intelligence unit, an annotation addition unit that acquires annotation information corresponding to the annotation target phrases output by the annotation target extraction unit and generates an annotated original text by adding the annotation information to the annotated target phrases, a translation unit that translates the annotated original text generated by the annotation addition unit into a language to be translated, and a display processing unit that displays the annotated translation translated by the translation unit on a display device.

[0007] A translation system according to the present disclosure includes the above-described translation device, and an artificial intelligence unit that, when an original text is input from the translation device, outputs to the translation device words and phrases to be annotated that are included in the original text.

[0008] The translation method according to the present disclosure includes: The computer an annotation target extraction step of inputting the original text acquired in the original text acquisition step into an artificial intelligence unit and acquiring and outputting annotation target phrases contained in the original text that are difficult to translate from the artificial intelligence unit; an annotation addition step of acquiring annotation information corresponding to the annotation target phrases output in the annotation target extraction step and generating an annotated original text by adding the annotation information to the original text; a translation step of generating an annotated translation by translating the annotated original text generated in the annotation addition step into a language to be translated; and a display control step of displaying the annotated translation obtained in the translation step on a display device. Execute. [Effects of the Invention]

[0009] According to the translation device, translation system, and translation method of the present disclosure, it is possible to display an easily understandable translation even when the original text contains words that are difficult to translate literally. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a functional block diagram illustrating a translation system according to a first embodiment of the present disclosure. [Figure 2]FIG. 2 is an explanatory diagram showing information stored in a user dictionary DB1. [Figure 3] 10A and 10B are explanatory diagrams illustrating examples of images that a display processing unit causes a display device to display; [Figure 4] 1 is a hardware configuration diagram illustrating a translation device according to a first embodiment of the present disclosure. [Figure 5] 3 is a flowchart showing a translation method according to the first embodiment of the present disclosure. [Figure 6] FIG. 2 is a functional block diagram showing a translation system according to a first modification of the first embodiment of the present disclosure. [Figure 7] FIG. 10 is a functional block diagram showing a translation system according to a second modification of the first embodiment of the present disclosure. [Figure 8] FIG. 10 is a functional block diagram showing a translation system according to a third modification of the first embodiment of the present disclosure. [Figure 9] FIG. 10 is a functional block diagram showing a translation system according to a fourth modification of the first embodiment of the present disclosure. [Figure 10] FIG. 10 is a functional block diagram showing a translation system according to a fifth modification of the first embodiment of the present disclosure. [Figure 11] FIG. 13 is a functional block diagram showing a translation system according to a sixth modification of the first embodiment of the present disclosure. [Figure 12] FIG. 12 is a functional block diagram showing a translation system according to a seventh modification of the first embodiment of the present disclosure. [Figure 13] FIG. 10 is a functional block diagram illustrating a translation system according to a second embodiment of the present disclosure. [Figure 14] FIG. 11 is a functional block diagram illustrating a translation system according to a third embodiment of the present disclosure. [Figure 15] FIG. 10 is a functional block diagram illustrating a translation system according to a fourth embodiment of the present disclosure. [Figure 16] 10 is a flowchart showing a translation method according to a fourth embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0011] An example of a translation device and a translation method according to the present disclosure will be described below with reference to the accompanying drawings. In the drawings, the same or corresponding parts are designated by the same reference numerals, and descriptions thereof will not be repeated.

[0012] Embodiment 1 <Functional configuration of the first embodiment> First, a translation system 1010 according to the first embodiment of the present disclosure will be described with reference to Fig. 1. Fig. 1 is a functional block diagram showing the translation system 1010.

[0013] The translation system 1010 is a system that uses artificial intelligence (AI) functions to annotate and translate phrases that are difficult to translate literally (hereinafter referred to as "phrases to be annotated"). Here, phrases to be annotated include, for example, abbreviations, buzzwords, technical terms, company jargon, or industry jargon. Phrases to be annotated are not limited to the above, and may be any phrases whose meaning may not be conveyed correctly even if translated directly. Technical terms are terms used in a specific field, company jargon is terms used within a specific team or company, and industry jargon is terms used in a specific industry or occupation.

[0014] First, an overview of each component of the translation system 1010 will be described. As shown in FIG. 1, the translation system 1010 includes a translation device 110, an artificial intelligence unit 200, a translation engine 300, and a display device 400.

[0015] Translation device 110 provides an original text to be translated to artificial intelligence unit 200, which has an artificial intelligence function, and acquires annotated phrases contained in the original text from artificial intelligence unit 200. Translation device 110 also adds annotations to the acquired annotated phrases. Translation device 110 then provides the annotated original text to translation engine 300, and causes display device 400 to display the annotated translation that translation engine 300 has translated into the target language.

[0016] When an original text is provided from translation device 110, artificial intelligence unit 200 extracts words and phrases to be annotated from the original text and outputs them to translation device 110. When an annotated original text is provided from translation device 110, translation engine 300 translates the annotated original text into the translation target language and outputs the resulting annotated translation to translation device 110. Display device 400 displays the annotated translation.

[0017] Next, the translation device 110 will be described. The translation device 110 includes an original text acquisition unit 10, an annotation target extraction unit 20, an annotation addition unit 30, a translation unit 40, a display processing unit 50, and a user dictionary DB.

[0018] The original text acquisition unit 10 acquires the original text to be translated. The original text acquisition unit 10 outputs the acquired original text to the annotation target extraction unit 20, the annotation addition unit 30, and the display processing unit 50.

[0019] The annotation target extraction unit 20 acquires the original text from the original text acquisition unit 10 . The annotation target extraction unit 20 inputs the original text acquired from the original text acquisition unit 10 to the artificial intelligence unit 200, and acquires annotation target phrases contained in the original text from the artificial intelligence unit 200. In detail, the annotation target extraction unit 20 causes the artificial intelligence unit 200 to translate the original text into the translation target language, and acquires from the artificial intelligence unit 200 phrases that were difficult for the artificial intelligence unit 200 to translate as annotation target phrases. The annotation target extraction unit 20 outputs the acquired annotation target words and phrases to the annotation adding unit 30 and the display processing unit 50.

[0020] The annotation adding unit 30 acquires the original text from the original text acquiring unit 10 and the annotation target phrases from the annotation target extracting unit 20 . The annotation adding unit 30 acquires annotation information corresponding to the annotation target phrases output by the annotation target extraction unit 20, and generates an annotated original text by adding the annotation information to the annotation target phrases included in the original text. In detail, the annotation adding unit 30 acquires annotation information corresponding to the annotation target phrases output by the annotation target extraction unit 20 from a user dictionary DB1 in which annotation target phrases registered in advance by a user are stored in association with annotation information. The annotation adding unit 30 outputs the generated annotated original text to the translation unit 40 and the display processing unit 50.

[0021] The user dictionary DB (Data Base) 1 is a storage unit that stores annotation information and associated words and phrases to be annotated that are registered in advance by a user. The user dictionary DB1 stores, as annotation information, at least one of the following: the definition of the word or phrase to be annotated, the reading of the word or phrase to be annotated, and an expression obtained by translating the definition of the word or phrase to be annotated into another language. The definition of the word or phrase to be annotated is an expression of the meaning or content of the word or phrase to be annotated in the same language as the original text. If the word or phrase to be annotated is an abbreviation, the definition of the word or phrase to be annotated includes the language before the abbreviation. The reading of the word or phrase to be annotated is an expression of the word or phrase to be annotated in phonetic characters. Here, for example, the same language as the original text is Japanese, and the other language is English.

[0022] Fig. 2 shows the information stored in the user dictionary DB1. As shown in Fig. 2, for example, the annotated phrase "Tei-tei-bi" (Regular Closing Day), which is an abbreviation and in-house term, is stored in association with the definition of the annotated phrase "Tei-ji-taisha-bi" (Regular Closing Day), the pronunciation of the annotated phrase "Teitaibi," and an expression obtained by translating the language of the annotated phrase into another language, "Regular Closing Day."

[0023] 2 shows an example in which, for one annotated phrase, the user dictionary DB1 stores multiple expressions obtained by translating the definition of the annotated phrase into multiple other languages, but this is not limiting. The user dictionary DB1 may store, for one annotated phrase, one expression corresponding to one other language.

[0024] 1 shows an example in which user dictionary DB1 is provided in translation device 110, but this is not limiting. User dictionary DB1 may be provided externally as long as it is configured to be connectable to translation device 110.

[0025] The translation unit 40 obtains the annotated original text from the annotation adding unit 30 . The translation unit 40 generates an annotated translation by translating the annotated original text generated by the annotation adding unit 30 into the translation target language. Specifically, the translation unit 40 provides the annotated original text to the translation engine 300 and obtains from the translation engine 300 an annotated translation by translating the annotated original text into the translation target language. The translation unit 40 outputs the generated annotated translation to the display processing unit 50.

[0026] The translation unit 40 may generate an annotated translation using the artificial intelligence unit 200. Specifically, the translation unit 40 may input an annotated original text to the artificial intelligence unit 200 and obtain from the artificial intelligence unit 200 an annotated translation that is the annotated original text translated into the translation target language.

[0027] The translation unit 40 causes the translation engine 300 or the artificial intelligence unit 200 to refer to the user dictionary DB1 and translate the annotated phrases and annotation information included in the annotated original text. Specifically, the translation unit 40 causes the translation engine 300 or the artificial intelligence unit 200 to translate the annotated phrases and annotation information using any one of the definition of the annotated phrase, the pronunciation of the annotated phrase, and an expression obtained by translating the definition of the annotated phrase into another language.

[0028] When the translation unit 40 translates the annotated phrase and the annotation information using the definition of the annotated phrase or an expression obtained by translating the definition of the annotated phrase into another language, the translation unit 40 translates the definition of the annotated phrase or the expression obtained by translating the definition of the annotated phrase into another language into the translation target language. When the translation unit 40 translates the annotated phrase and the annotation information using an expression obtained by translating the definition of the annotated phrase into another language and if the translation target language and the other language are the same, the translation unit 40 uses the expression obtained by translating the definition of the annotated phrase into the other language as is. When the translation unit 40 translates the annotated phrase and the annotation information using the pronunciation of the annotated phrase, the translation unit 40 uses the pronunciation of the annotated phrase as is, or translates the pronunciation of the annotated phrase into phonetic characters corresponding to the translation target phrase.

[0029] Furthermore, when the translation unit 40 translates a phrase to be annotated using any one of the definition of the phrase to be annotated, the reading of the phrase to be annotated, and an expression obtained by translating the definition of the phrase to be annotated into another language, it is desirable to use different expressions for the translation of the phrase to be annotated and the expression used as annotation information. For example, when the definition of the phrase to be annotated is used as annotation information, it is desirable to use the reading of the phrase to be annotated when translating the phrase to be annotated.

[0030] In the above, an example has been described in which translation unit 40 performs translation by having translation engine 300 or artificial intelligence unit 200 refer to user dictionary DB1, but this is not limiting. Even if translation unit 40 translates a word or phrase to be annotated without referring to user dictionary DB1, resulting in a literal translation of the word or phrase to be annotated that does not correctly convey the meaning, an easily understandable translation can be displayed using the literal translation annotation information.

[0031] The display processing unit 50 acquires original text from the original text acquisition unit 10, annotation target phrases from the annotation target extraction unit 20, annotated original text from the annotation adding unit 30, and annotated translation text from the translation unit . The display processing unit 50 displays the original text acquired by the original text acquisition unit 10, the annotated original text generated by the annotation adding unit 30, and the annotated translation generated by the translation unit 40 on the display device 400.

[0032] The display device 400, whose display is controlled by the display processing unit 50, includes an audience terminal for the audience and an operation terminal for the user. Here, an audience refers to a person who checks the annotated translation to understand the content of the original text, and a user refers to a person who operates the content displayed on the audience terminal. The display processing unit 50 causes the audience terminal to display, for example, the original text and the annotated translation. The display processing unit 50 causes the operation terminal to display, for example, the original text, the annotated original text, and the annotated translation.

[0033] 3 shows examples of images that the display processing unit 50 displays on the display device 400. Note that Fig. 3(a) is an example of an image that is displayed on the auditory terminal, and Fig. 3(b) is an example of an image that is displayed on the operating terminal.

[0034] As shown in Figure 3(a), the display processing unit 50 displays the original text and the annotated translation side by side on the listener terminal. Listeners can easily understand the content of the original text by checking the annotation information displayed on the listener terminal.

[0035] 3(a) shows an example in which the display processing unit 50 displays an original text and multiple annotated translations corresponding to multiple target languages ​​side by side. By displaying multiple annotated translations side by side, the display processing unit 50 can simultaneously accommodate multiple listeners who use multiple target languages. The display processing unit 50 may display only the annotated translation on the listening terminal. Also, the display processing unit 50 may display only one annotated translation corresponding to one phrase to be annotated.

[0036] As shown in Figure 3(b), the display processing unit 50 displays the original text, the annotated original text, and the annotated translation on the operating terminal. For example, the display processing unit 50 displays the original text or the annotated original text on the left side of the operating terminal screen, and the annotated original text on the right side. By checking the original text, the annotated original text, and the annotated translation displayed on the operating terminal, the user can determine whether the annotated words and phrases and annotation information contained in the original text and the annotated translation to be displayed on the listening terminal are appropriate.

[0037] 3(b) shows an example in which the display processing unit 50 displays the annotated original text and the annotated translation on the operation terminal so that the annotated phrases and annotation information are highlighted with markers. By displaying the annotated phrases and annotation information on the operation terminal so that the display processing unit 50 highlights them, the user can easily determine whether the annotated phrases and annotation information displayed on the listening terminal are appropriate. The display processing unit 50 may display the annotated translation on the display terminal so as to highlight the words and phrases to be annotated and the annotation information. The highlighting method is not limited to highlighting, and may be bold, colored, underlined, or the like.

[0038] 3 shows an example in which the display processing unit 50 displays annotation information included in the annotated translation in parentheses at the end of the annotated phrase. Note that the display of annotation information by the display processing unit 50 is not limited to this. The display processing unit 50 may display the annotation information included in the annotated translation in a way that makes it clear how it corresponds to the annotated phrase. For example, the display processing unit 50 may mark the annotated phrase and then display the annotation information at the end of the sentence with the same mark.

[0039] Next, the artificial intelligence unit 200 will be described. The artificial intelligence unit 200 is an artificial intelligence equipped with intelligent functions such as inference and judgment, and its operating environment. When an original text is input from the translation device 110, the artificial intelligence unit 200 outputs words and phrases to be annotated contained in the original text based on the original text and a trained model (described later). The artificial intelligence unit 200 performs one of morphological analysis, syntactic interpretation, semantic analysis, and contextual analysis to extract words and phrases to be annotated. As shown in FIG. 1, the artificial intelligence unit 200 includes a trained model storage unit 201 and a model control unit 202.

[0040] The trained model storage unit 201 stores a trained model that has been trained to output annotation target phrases contained in an original sentence when the original sentence is input. For example, a large language model (LLM) is used as the trained model. By inputting the original sentence as an analysis target and a prompt including a description instructing the LLM to output annotation target phrases contained in the original sentence, it is possible to extract annotation target phrases contained in the original sentence.

[0041] The trained model is not limited to the LLM. The trained model includes model information, which will be described later. The trained model may include model parameters, which are information that defines the behavior of the model, such as constraints, weighting variables, and evaluation functions. The model may be, for example, a model called a neural network (NN), a convolutional neural network (CNN), a recurrent neural network (RNN), a variational autoencoder (VAE), a generative adversarial network (GAN), a diffusion model, a transformer, a visual language model (VLM), a bidirectional encoder representations from transformers (BERT), a generative pre-trained transformer (GPT), or a contrastive language image pre-training (CLIP). Note that the above-mentioned models are not exclusive; for example, LLM, VLM, BERT, and GPT are included in the transformer. Also, for example, the transformer is included in the neural network. Furthermore, the learning algorithm and model may be a combination of multiple types. The model also includes a multimodal model that is trained by combining multiple different types of data.

[0042] When the model control unit 202 acquires an original text from the translation device 110, it extracts words and phrases to be annotated that are included in the original text based on the original text and the trained model, and outputs the extracted words and phrases to the translation device 110. That is, when the model control unit 202 acquires an original text, it extracts words and phrases to be annotated that are included in the original text using the model indicated by the trained model.

[0043] Specifically, the model control unit 202 inputs an original sentence as an analysis target and a prompt including a statement instructing the LLM, which is a trained model, to output an annotation target phrase contained in the original sentence, and obtains the analysis processing result from the LLM. The model control unit 202 obtains the original sentence or the prompt from the annotation target extraction unit 20 of the translation device 110. That is, the model control unit 202 obtains at least the original sentence contained in the prompt from the annotation target extraction unit 20 of the translation device 110. The prompt may be acquired by the translation device 110 receiving input, or may be automatically generated using a stored prompt template. If the prompt is automatically generated, the automatic generation of the prompt may be performed by the annotation target extraction unit 20 of the translation device 110 or the model control unit 202. The prompt template may be stored in the translation device 110 or the artificial intelligence unit 200. The annotation target extraction unit 20 of the translation device 110 or the model control unit 202 reads out a prompt template stored in the translation device 110 or the artificial intelligence unit 200, and automatically generates a prompt by embedding a description of the original text to be analyzed into the read template.

[0044] The instruction to annotate the phrases included in the prompt may be, for example, an instruction to translate the original text and extract the difficult-to-translate phrases. The LLM receives the prompt, translates the original text, and extracts the difficult-to-translate phrases. The model control unit 202 outputs the difficult-to-translate phrases extracted by the LLM as annotate phrases to the annotation target extraction unit 20 of the translation device 110. Here, the difficult-to-translate phrases are words or phrases that do not fit the context, words or phrases that do not correspond to common words or phrases known by the large-scale language model, words or phrases that are frequently mistranslated and stored in the artificial intelligence unit 200, or words or phrases with a low confidence score generated by the artificial intelligence unit 200.

[0045] The trained model and other information used by the artificial intelligence unit 200 may be prepared in advance, or may be acquired via a network as needed.

[0046] 1 shows an example in which the artificial intelligence unit 200 is configured externally to the translation device 110. For example, the artificial intelligence unit 200 is an external computer such as a cloud computer, and the translation device 110 is configured to be connectable to the external computer through API (Application Programming Interface) cooperation or the like. Note that the translation device 110 may be provided with the artificial intelligence unit 200, or the translation device 110 may use the artificial intelligence unit 200 that the translation device 110 itself has.

[0047] Next, the translation engine 300 will be described. Translation engine 300 is a device that translates a sentence provided by translation device 110 and outputs the translation result to translation device 110. In detail, translation engine 300 acquires an annotated original sentence from translation unit 40 of translation device 110, and outputs an annotated translation obtained by translating the annotated original sentence to translation unit 40 of translation device 110. Translation engine 300 performs translation using, for example, a rule-based algorithm. 1 shows an example in which the translation engine 300 is provided outside the translation device 110. However, this is merely an example, and the translation engine 300 may be provided inside the translation device 110. For example, if the translation unit 40 is provided with a translation engine, the translation unit 40 translates the original text to be translated into the translation target language.

[0048] 1 shows an example in which the translation unit 40 of the translation device 110 uses the translation engine 300 to perform the translation, but this is not limiting. As described above, the translation unit 40 only needs to be able to translate the annotated original text, and may also use the artificial intelligence unit 200 to perform the translation.

[0049] The display device 400 will now be described. Display device 400 is a device that displays annotated translations and the like obtained from translation device 110. Display device 400 obtains the original text, annotated original text, and annotated translations from display processing unit 50 of translation device 110 and displays them.

[0050] As described above, the display device 400 includes an audience terminal for the audience and an operation terminal for the user. For example, the audience terminal displays the original text and the annotated translation, and the operation terminal displays the original text, the annotated original text, and the annotated translation.

[0051] <Hardware Configuration of First Embodiment> Next, the hardware configuration of translation device 110 according to the first embodiment will be described with reference to Fig. 4. Fig. 4 is a diagram showing the hardware configuration of translation device 110.

[0052] 4, translation device 110 is configured with processor 101, auxiliary storage device 102, storage device 103, and input / output interface 104. Processor 101, auxiliary storage device 102, storage device 103, and input / output interface 104 are connected via signal line 105.

[0053] The processor 101 is a device that realizes the functions of the original text acquisition unit 10, the annotation target extraction unit 20, the annotation addition unit 30, the translation unit 40, and the display processing unit 50. The processor 101 reads out necessary programs from the auxiliary storage device 102 and executes the processes, thereby realizing the functions of the original text acquisition unit 10, the annotation target extraction unit 20, the annotation addition unit 30, the translation unit 40, and the display processing unit 50 of the translation device 110. The processor 101 is, for example, an integrated circuit (IC) that performs arithmetic processing, such as a central processing unit (CPU), a digital signal processor (DSP), or a graphics processing unit (GPU).

[0054] 1, and is an auxiliary storage device of the translation device 110. The auxiliary storage device 102 stores translation programs required to realize the functions of the translation device 110, words and phrases to be annotated, and annotation information corresponding to the words and phrases to be annotated. The auxiliary storage device 102 is, for example, a read-only memory (ROM), a hard disk drive (HDD), or a solid state drive (SSD).

[0055] The storage device 103 is a main storage device of the translation device 110. The main storage device temporarily stores calculations for processing performed by the processor 101. The storage device 103 is, for example, a RAM (Random Access Memory).

[0056] The input / output interface 104 is an interface for inputting and outputting data to and from an external device. The input / output interface 104 inputs and outputs data between the artificial intelligence unit 200, the translation engine 300, and the display device 400. The input / output interface 104 is, for example, an Ethernet (registered trademark), USB (Universal Serial Bus), or HDMI (registered trademark, High-Definition Multimedia Interface) port.

[0057] The signal line 105 is a transmission path for transmitting and receiving data between the components shown in FIG.

[0058] In the above description, the translation device 110, the artificial intelligence unit 200, the translation engine 300, and the display device 400 are configured as independent pieces of hardware and interconnected, but this is not limiting. At least some of the above devices may be configured as an integrated unit. That is, at least some of the functional blocks shown in FIG. 1 may be realized by hardware, or may be configured to be realized by a combination of hardware and software.

[0059] <Translation method according to the first embodiment> Next, the procedure of the translation method according to the first embodiment will be described with reference to Fig. 5. Fig. 5 is a flowchart showing the translation method according to the first embodiment. Note that the following describes an example in which an original Japanese text is translated into English.

[0060] First, the original text acquisition unit 10 acquires an original text in Japanese as an original text to be translated (step ST1). For the sake of convenience, it is assumed here that the original Japanese text acquired is "Today is your final day." Furthermore, the original text acquisition unit 10 outputs the original text to the annotation target extraction unit 20, the annotation addition unit 30, and the display processing unit 50. The display processing unit 50 causes the display device 400 to display the original text output from the original text acquisition unit 10.

[0061] Next, the annotation target extraction unit 20 inputs the original text acquired by the original text acquisition unit 10 to the artificial intelligence unit 200 (step ST2). The annotation target extraction unit 20 acquires, from the artificial intelligence unit 200, annotation target phrases that are difficult to translate and are included in the original text acquired by the original text acquisition unit 10. When the annotation target extraction unit 20 acquires the annotation target phrases from the artificial intelligence unit 200 (YES in step ST3), the annotation target extraction unit 20 outputs the annotation target phrases acquired from the artificial intelligence unit 200 to the annotation adding unit 30 (step ST4). When the annotation target extraction unit 20 does not acquire the annotation target phrases from the artificial intelligence unit 200 (NO in step ST3), the annotation target extraction unit 20 does not output the annotation target phrases to the annotation adding unit 30 and translates the original text (step ST9).

[0062] Specifically, the annotation target extraction unit 20 causes the artificial intelligence unit 200 to translate the original text acquired by the original text acquisition unit 10 into the translation target language and extracts words and phrases that were difficult to translate by the artificial intelligence unit 200. When words and phrases that were difficult to translate by the artificial intelligence unit 200 are extracted, the annotation target extraction unit 20 acquires the extracted words and phrases that were difficult to translate by the artificial intelligence unit 200 from the artificial intelligence unit 200 as words and phrases to be annotated (YES in step ST3). When words and phrases that were difficult to translate by the artificial intelligence unit 200 are not extracted, the annotation target extraction unit 20 does not acquire words and phrases to be annotated from the artificial intelligence unit 200 (NO in step ST3).

[0063] In other words, the annotation target extraction unit 20 inputs the original text acquired by the original text acquisition unit 10, a description instructing the original text to be translated, and a description instructing the extraction of words that were difficult to translate during translation into the artificial intelligence unit 200, and acquires from the artificial intelligence unit 200 the words that were difficult to translate and that the artificial intelligence unit 200 extracted based on the input as annotation target words. Here, words or phrases that were difficult for the artificial intelligence unit 200 to translate are words or phrases that do not fit the context, words or phrases that are not stored in the standard dictionary DB 203 that stores common words or phrases that the artificial intelligence unit 200 refers to when translating, words or phrases that are frequently mistranslated and that are stored in the artificial intelligence unit 200, or words or phrases that have a low reliability score generated by the artificial intelligence unit 200.

[0064] For example, when the annotation target extraction unit 20 inputs "Today is the closing day" to the artificial intelligence unit 200, "closing day" is extracted as a phrase that was difficult for the artificial intelligence unit 200 to translate, and the annotation target extraction unit 20 obtains "closing day" from the artificial intelligence unit 200 as a phrase to be annotated.

[0065] Furthermore, the annotation target extraction unit 20 outputs the annotation target phrases acquired from the artificial intelligence unit 200 to the display processing unit 50. The display processing unit 50 displays the annotation target phrases output from the annotation target extraction unit 20 on the display device 400 so as to be highlighted. For example, as shown in FIG. 3, the display processing unit 50 displays a marker symbol on the display portion of "final day off" so as to highlight the annotation target phrase "final day off" contained in the original text "Today is your final day off."

[0066] The annotation adding unit 30 acquires annotation information corresponding to the annotated phrase output by the annotation target extraction unit 20 (YES in step ST5). When the annotation adding unit 30 acquires annotation information corresponding to the annotated phrase (YES in step ST5), it generates an annotated original text by adding the annotation information to the annotated phrase (step ST6). When the annotation adding unit 30 does not acquire annotation information corresponding to the annotated phrase (NO in step ST5), it does not generate an annotated original text and translates the original text (step ST9).

[0067] Specifically, the annotation adding unit 30 refers to a user dictionary DB1 in which annotation target words and phrases registered in advance by the user are stored in association with annotation information, and acquires annotation information corresponding to the annotation target words and phrases output by the annotation target extraction unit 20. If the user dictionary DB1 stores annotation information corresponding to the annotation target words and phrases output by the annotation target extraction unit 20, the annotation adding unit 30 acquires the annotation information stored in the user dictionary DB1 (YES in step ST5). If the user dictionary DB1 does not store annotation information corresponding to the annotation target words and phrases output by the annotation target extraction unit 20, the annotation adding unit 30 does not acquire the annotation information (NO in step ST5).

[0068] For example, the annotation adding unit 30 refers to the user dictionary DB1 to check whether annotation information corresponding to the annotation target phrases "regular leaving day" and "regular leaving day" output by the annotation target extraction unit 20 is stored. If the annotation information "regular leaving day" corresponding to "regular leaving day" is stored in the user dictionary DB1, the annotation adding unit 30 acquires "regular leaving day" from the user dictionary DB1 as the annotation information corresponding to "regular leaving day" and generates an annotated original sentence by adding the annotation information "regular leaving day" to "regular leaving day". If the annotation information corresponding to "regular leaving day" is not stored in the user dictionary DB1, the annotation adding unit 30 does not acquire annotation information corresponding to "regular leaving day" and does not generate an annotated original sentence.

[0069] The annotation adding unit 30 outputs the generated annotated original text to the translation unit 40. The annotation adding unit 30 also outputs the generated annotated original text to the display processing unit 50. The display processing unit 50 displays the annotated original text "Today is the regular closing day (the day when you leave work on time)." in which the annotation target phrase and the annotation target phrase are associated with each other. For example, as shown in FIG. 3, the display processing unit 50 displays the annotated original text "Today is the regular closing day (the day when you leave work on time)." in which the annotation information "the day when you leave work on time" is added in parentheses to the end of the annotation target phrase "the day when you leave work on time" contained in the original text "Today is the regular closing day."

[0070] The translation unit 40 generates an annotated translation by translating the annotated original text generated by the annotation adding unit 30 into English, which is the target language for translation (step ST7). Specifically, the annotated original text is provided to the translation engine 300, and an annotated English translation is obtained from the translation engine 300. The translation unit 40 outputs the annotated translation to the display processing unit 50.

[0071] The display processing unit 50 displays the annotated translation produced by the translation unit 40 on the display device (step ST8).

[0072] Furthermore, if the annotation adding unit 30 does not acquire annotation information corresponding to the annotated phrase (NO in step ST5), or if the annotated phrase is not acquired from the artificial intelligence unit 200 (NO in step ST3), the translation unit 40 generates a translation by translating the original text acquired by the original text acquisition unit 10 into English, which is the translation target language (step ST9).Then, the display processing unit 50 displays the translation produced by the translation unit 40 on the display device (step ST10).

[0073] <Effects of the First Embodiment> The following describes the operations and effects of translation device 110, translation system 1010, and translation method according to the first embodiment of the present disclosure.

[0074] The translation device 110 according to the first embodiment of the present disclosure includes an original text acquisition unit 10 that acquires an original text to be translated, an annotation target extraction unit 20 that inputs the original text acquired by the original text acquisition unit 10 to an artificial intelligence unit 200 and acquires and outputs annotation target phrases contained in the original text that are difficult to translate from the artificial intelligence unit 200, an annotation addition unit 30 that acquires annotation information corresponding to the annotation target phrases output by the annotation target extraction unit 20 and generates an annotated original text by adding the annotation information to the annotated target phrases, a translation unit 40 that translates the annotated original text generated by the annotation addition unit 30 into a language to be translated, and a display processing unit 50 that displays the annotated translation translated by the translation unit 40 on a display device 400.

[0075] The translation system 1010 according to the first embodiment of the present disclosure includes the above-described translation device 110 and an artificial intelligence unit 200 that, when an original text is input from the translation device 110, outputs to the translation device 110 words to be annotated that are included in the original text.

[0076] The translation method according to the first embodiment of the present disclosure includes an original text acquisition step of acquiring an original text to be translated, an annotation target extraction step of inputting the original text acquired in the original text acquisition step into the artificial intelligence unit 200 and acquiring and outputting annotation target phrases contained in the original text that are difficult to translate from the artificial intelligence unit 200, an annotation addition step of acquiring annotation information corresponding to the annotation target phrases output in the annotation target extraction step and generating an annotated original text by adding the annotation information to the original text, a translation step of generating an annotated translation by translating the annotated original text generated in the annotation addition step into a language to be translated, and a display control step of displaying the annotated translation obtained in the translation step on the display device 400.

[0077] According to the translation device 110, the translation system 1010, and the translation method according to the first embodiment of the present disclosure, it is possible to annotate phrases that are difficult to translate literally. Therefore, even if the original text contains phrases that are difficult to translate literally, it is possible to display an easily understandable translation.

[0078] In addition, the annotation target extraction unit 20 provided in the translation device 110 according to the first embodiment of the present disclosure causes the artificial intelligence unit 200 to translate the original text into the target language, and obtains and outputs from the artificial intelligence unit 200 the words and phrases that were difficult for the artificial intelligence unit 200 to translate as words and phrases to be annotated. According to the translation device 110, it is possible to automatically extract from the original text words that are difficult to translate, which correspond to words that are difficult to translate literally, by the artificial intelligence unit 200. Therefore, it is possible to reduce the burden on the user and display an easily understandable translation even when the original text contains words that are difficult to translate literally.

[0079] The annotation adding unit 30 included in the translation device 110 according to the first embodiment of the present disclosure acquires annotation information corresponding to the annotated phrase output by the annotation target extraction unit 20 from a memory unit (user dictionary DB1) in which annotation information is stored in association with annotated phrases registered in advance by the user. According to the translation device 110, the user can display the annotations intended by the user by registering annotation information corresponding to the words and phrases to be annotated in advance. Therefore, even if the original text contains words and phrases that are difficult to translate literally, a translation that more appropriately reflects the meaning of the words and phrases that the user wants to convey can be displayed.

[0080] The annotation adding unit 30 included in the translation device 110 according to the first embodiment of the present disclosure adds at least one of the definition of the phrase to be annotated, the pronunciation of the phrase to be annotated, and an expression obtained by translating the definition of the phrase to be annotated into another language as annotation information to the phrase to be annotated. According to the above-mentioned translation device 110, at least one of the following can be displayed as annotations for the phrase to be annotated: the definition of the phrase to be annotated, the pronunciation of the phrase to be annotated, and an expression obtained by translating the definition of the phrase to be annotated into another language.

[0081] The translation unit 40 included in the translation device 110 according to the first embodiment of the present disclosure translates the phrase to be annotated using one of the definition of the phrase to be annotated, the reading of the phrase to be annotated, and an expression obtained by translating the definition of the phrase to be annotated into another language. According to the translation device 110, it is possible to translate an annotated phrase that is difficult to translate literally into content that represents at least one of the definition of the annotated phrase, the pronunciation of the annotated phrase, and an expression obtained by translating the definition of the annotated phrase into another language, without directly translating the annotated phrase. Therefore, even if the original text contains a phrase that is difficult to translate literally, the phrase itself that is difficult to translate literally can be translated into an easily understandable expression and displayed.

[0082] The display processing unit 50 included in the translation device 110 according to the first embodiment of the present disclosure further displays the annotated original text so as to highlight the words and phrases to be annotated. The translation device 110 described above makes it easier for a user or listener checking the display device to check the annotated phrases contained in the original text. For example, it has the effect that a user who operates the content to be displayed on the listening terminal can easily determine whether the annotated phrases to be displayed on the listening terminal are appropriate.

[0083] Variation 1. In the first embodiment, a translation device 110 has been described that includes an annotation target extraction unit 20 that acquires, as annotated phrases, phrases that are difficult to translate by the artificial intelligence unit 200. In the first variation of the first embodiment, a translation device 111 has an annotation target extraction unit 21 that receives a specification of annotated phrases from a user, and acquires and outputs annotated phrases corresponding to the received specification from the artificial intelligence unit 200.

[0084] A translation device 111 according to a first modification of the first embodiment will be described with reference to Fig. 6. Fig. 6 is a functional block diagram showing a translation system 1011 including the translation device 111.

[0085] As shown in FIG. 6, translation device 111 according to variant 1 of embodiment 1 differs from translation device 110 according to embodiment 1 in that it further includes a phrase specification receiving unit 60 and an annotation target extraction unit 21 that is different from that of embodiment 1.

[0086] The phrase designation receiving unit 60 receives designation of the phrase to be annotated from the user via the operation terminal. Specifically, the phrase designation receiving unit 60 causes the operation terminal to display an image for accepting the designation of the phrase to be annotated. The user designates the phrase in accordance with the display on the operation terminal. The phrase designation receiving unit 60 acquires the designation of the phrase to be annotated received by the operation terminal. For example, the phrase specification receiving unit 60 receives input of a description specifying a phrase to be annotated, such as "Please add an annotation to the word 'regular retirement date'."

[0087] The annotation target extraction unit 21 acquires from the artificial intelligence unit 200 annotation target phrases that are included in the original text and that correspond to the specifications received by the phrase specification receiving unit 60, and outputs them. For example, assume that the phrase specification receiving unit 60 acquires a description specifying the phrase to be annotated, "Please annotate the word 'regular retirement date,'" and the original text acquisition unit 10 acquires the original text, "Today is the regular retirement date." In this case, the annotation target extraction unit 22 inputs the description specifying the phrase to be annotated received by the phrase specification receiving unit 60 and the original text acquired by the original text acquisition unit 10 to the artificial intelligence unit 200, and acquires and outputs the phrase to be annotated, "regular retirement date," included in the original text, from the artificial intelligence unit 200 in response to the specification received by the phrase specification receiving unit 60.

[0088] Furthermore, the annotated phrase corresponding to the specification received by the phrase specification receiving unit 60 may include not only the same phrase as the specified phrase, but also phrases determined by the artificial intelligence unit 200 to be substantially identical to the specified phrase due to spelling variations or other reasons. For example, if the user specifies "Date of Retirement" (the "Date of Retirement"), a description specifying the annotated phrase may read, "Please annotate Date of Retirement or words determined to have the same meaning as it." The annotated phrase corresponding to the received specification may exclude phrases that the artificial intelligence unit 200 determines to be substantially different from the specified phrase, even if they contain the same character string as the specified phrase. For example, if the user specifies "GL," an abbreviation for "group leader," the characters "GL" included in the phrase "GLOBAL" in the original text are not extracted as an annotated phrase. In this case, a description specifying the annotated phrase may be, "Please annotate the word GL, but exclude words determined to be synonyms based on the context."

[0089] The annotation target extraction unit 21 may replace the annotation target extraction unit 20 according to the first embodiment, or may perform extraction by designation in addition to the functions of the annotation target extraction unit 20 according to the first embodiment. That is, instead of acquiring from the artificial intelligence unit 200, words that were difficult to translate by the artificial intelligence unit 200 as annotation target words, the annotation target extraction unit 21 may acquire, from the artificial intelligence unit 200, annotation target words corresponding to the designation accepted by the phrase designation accepting unit 60, or may acquire, from the artificial intelligence unit 200, words that were difficult to translate by the artificial intelligence unit 200 and annotation target words corresponding to the designation accepted by the phrase designation accepting unit 60 as annotation target words. Furthermore, after acquiring words that were difficult to translate by the artificial intelligence unit 200, the annotation target extraction unit 21 may accept designation of words from among words that were difficult to translate by the artificial intelligence unit 200 and acquire annotation target words corresponding to the designation.

[0090] 6, or before or after ST3, the annotation target extraction unit 21 acquires and outputs from the artificial intelligence unit 200 annotated phrases included in the original text in response to a specification accepted by the phrase specification accepting unit 60. When acquiring annotated phrases corresponding to a specification after ST3, the phrase specification accepting unit 60 may accept a specification of annotated phrases from among the phrases that were difficult to translate by the artificial intelligence unit 200 and that were acquired in ST3, and the annotation target extraction unit 21 may acquire from the artificial intelligence unit 200 annotated phrases corresponding to the specification.

[0091] The following describes the operations and effects of translation device 111, translation system 1011, and translation method according to Variation 1 of Embodiment 1 of the present disclosure. Translation device 111 according to Variation 1 of Embodiment 1 of the present disclosure further includes a phrase designation receiving unit 60 that receives designation of phrases to be annotated. Additionally, annotation target extraction unit 21 included in translation device 111 acquires, from artificial intelligence unit 200, an annotation target phrase included in the original text that corresponds to the designation received by phrase designation receiving unit 60, and outputs the annotated phrase. According to translation device 111 according to Variation 1 of Embodiment 1 of the present disclosure, a user can specify a phrase to which the user wants to add annotation information. Therefore, even if the original text contains a phrase that is difficult to translate directly, the user can add an annotation to the phrase to which the user wants to add an annotation, and display a translation that is easier to understand.

[0092] Variation 2. In the first variation of the first embodiment, a translation device 111 is described that includes an annotation target extraction unit 21 that extracts annotated phrases specified by a user. In the second variation of the first embodiment, a translation device 112 is described that includes an annotation target extraction unit 22 that refers to a user dictionary DB1, acquires annotated phrases stored in the user dictionary DB1 that are included in an original text from an artificial intelligence unit 200, and outputs the annotated phrases.

[0093] A translation device 112 according to the second modification of the first embodiment will be described with reference to Fig. 7. Fig. 7 is a functional block diagram showing a translation system 1012 including the translation device 112.

[0094] As shown in FIG. 7, translation device 112 according to the second modification of the first embodiment differs from translation device 111 according to the first modification of the first embodiment in the annotation target extraction unit 22.

[0095] Like translation device 110 according to embodiment 1, translation device 112 includes a user dictionary DB1 that stores annotated phrases registered in advance by a user. Annotation target extraction unit 22 included in translation device 112 causes artificial intelligence unit 200 to refer to user dictionary DB1, and acquires and outputs annotated phrases that are included in the original text and stored in user dictionary DB1 from the artificial intelligence unit. For example, assume that "Date to Leave Work" is stored in user dictionary DB1 as shown in Fig. 2, and the original sentence obtained is "Today is the date to leave work." In this case, annotation target extraction unit 22 inputs a statement instructing to refer to user dictionary DB1 and the original sentence to artificial intelligence unit 200, and obtains and outputs from artificial intelligence unit 200 the annotation target phrase "Date to Leave Work" that is included in the original sentence and stored in user dictionary DB1.

[0096] In addition, the annotation target extraction unit 22 may replace the annotation target extraction unit 20 according to embodiment 1 and the annotation target extraction unit 21 according to variant 1 of embodiment 1, or may extract annotation target words and phrases stored in the user dictionary DB1 in addition to the functions of at least one of the annotation target extraction unit 20 according to embodiment 1 and the annotation target extraction unit 21 according to variant 1 of embodiment 1.

[0097] In a translation method using the translation device 112 according to the second variant of the first embodiment, instead of ST3 shown in FIG. 5, or before or after ST3, the annotation target extraction unit 23 causes the artificial intelligence unit 200 to refer to the user dictionary DB1, and obtains and outputs the annotation target words and phrases stored in the user dictionary DB1 that are included in the original text from the artificial intelligence unit 200.

[0098] The following describes the operations and effects of the translation device 112, the translation system 1012, and the translation method according to the second modification of the first embodiment of the present disclosure. The annotation target extraction unit 22 provided in the translation device 112 according to the second variation of the first embodiment of the present disclosure causes the artificial intelligence unit 200 to refer to a memory unit (user dictionary DB1) that stores annotation target words and phrases that have been registered in advance by the user, and obtains and outputs the annotation target words and phrases stored in the memory unit (user dictionary DB1) that are included in the original text from the artificial intelligence unit 200. According to the translation device 112 according to the second modification of the first embodiment of the present disclosure, the user can register in advance the words and phrases to which he / she wants to add annotation information, thereby extracting the words and phrases to be annotated. Therefore, even if the original text contains words and phrases that are difficult to translate literally, the user can add annotations to the words and phrases to which he / she wants to add annotations, and display a translation that is easier to understand.

[0099] Variation 3. In the second variation of the first embodiment, we have described a translation device 112 equipped with an annotation target extraction unit 22 that refers to a user dictionary DB1, acquires annotated phrases stored in the user dictionary DB1 that are included in an original text from the artificial intelligence unit 200, and outputs the acquired phrases. In the third variation of the first embodiment, we will describe a translation device 113 equipped with an annotation target extraction unit 23 that refers to a history information DB2 that stores annotated phrases included in an original text that have been acquired in the past, and acquires annotated phrases stored in the history information DB2 that are included in the original text, and outputs the acquired phrases.

[0100] A translation device 113 according to a third modification of the first embodiment will be described with reference to Fig. 8. Fig. 8 is a functional block diagram showing a translation system 1013 including the translation device 113.

[0101] As shown in FIG. 8, translation device 113 according to variant example 3 of embodiment 1 differs from translation device 112 according to variant example 2 of embodiment 1 in that it further includes a history information DB2 and an annotation target extraction unit 23 and an annotation addition unit 31 that are different from those in variant example 2 of embodiment 1.

[0102] The historical information DB2 is a memory unit that stores, in association with each other, words to be annotated contained in original texts acquired in the past (hereinafter referred to as past words to be annotated) and annotation information corresponding to the past words to be annotated (hereinafter referred to as past annotation information). For example, the historical information DB2 stores the past annotation target phrase "regular closing day" contained in the previously acquired original sentence "Tomorrow is the regular closing day" in association with the past annotation information "day to leave work on time."

[0103] The annotation target extraction unit 23 causes the artificial intelligence unit 200 to refer to the history information DB2, and acquires from the artificial intelligence unit and outputs the annotation target words and phrases contained in the original text and stored in the history information DB2. For example, assume that the historical information DB2 stores a past annotation target phrase "regular leaving day" and past annotation information "date of leaving work on time" in association with each other, and that the original text "Today is the regular leaving day" is acquired. In this case, the annotation target extraction unit 23 inputs a statement instructing the AI ​​unit 200 to refer to the historical information DB2 and the original text, and acquires and outputs the annotation target phrase "regular leaving day" included in the original text and stored in the historical information DB2 from the AI ​​unit 200.

[0104] The annotation adding unit 31 acquires annotation information corresponding to the annotation target phrase output by the annotation target extracting unit 23 from the history information DB2. For example, assume that the historical information DB2 stores a past annotated phrase "regular leave date" and past annotation information "date of leaving work on time" in association with each other, and that the annotation target extraction unit 23 outputs "regular leave date" as the current annotated phrase. In this case, the annotation adding unit 31 obtains the annotation information "date of leaving work on time" corresponding to the annotation target phrase "regular leave date" output by the annotation target extraction unit 23 from the historical information DB2.

[0105] The annotation target extraction unit 23 may replace the annotation target extraction unit 20 according to embodiment 1, the annotation target extraction unit 21 according to variant 1 of embodiment 1, and the annotation target extraction unit 22 according to variant 2 of embodiment 1, or may extract annotation target terms stored in the history information DB2 in addition to performing at least one of the functions of the annotation target extraction unit 20 according to embodiment 1, the annotation target extraction unit 21 according to variant 1 of embodiment 1, and the annotation target extraction unit 22 according to variant 2 of embodiment 1. Furthermore, the annotation addition unit 31 may replace the annotation addition unit 30 according to embodiment 1, or may acquire annotation information stored in the history information DB2 in addition to the functions of the annotation addition unit 30 according to embodiment 1.

[0106] In a translation method using translation device 113 according to Variation 3 of Embodiment 1, instead of ST3 shown in Fig. 5, or before or after ST3, annotation target extraction unit 23 causes artificial intelligence unit 200 to refer to history information DB2, and acquires and outputs annotated phrases contained in the original text and stored in history information DB2 from artificial intelligence unit 200. Also, instead of ST5 shown in Fig. 5, or before or after ST5, annotation adding unit 31 checks whether annotation information corresponding to the annotated phrases output by annotation target extraction unit 23 is stored in history information DB2.

[0107] The following describes the operations and effects of the translation device 113, the translation system 1013, and the translation method according to the third modification of the first embodiment of the present disclosure. An annotation target extraction unit 23 included in translation device 113 according to Variation 3 of Embodiment 1 of the present disclosure causes artificial intelligence unit 200 to refer to a memory unit (history information DB2) in which previously acquired original sentences and annotation target phrases included in the previously acquired original sentences are stored in association with each other, and acquires and outputs the annotation target phrases included in the original sentences stored in the memory unit (history information DB2) from artificial intelligence unit 200. Furthermore, annotation adding unit 31 included in translation device 113 acquires annotation information corresponding to the annotation target phrases output by annotation target extraction unit 23 from the memory unit (history information DB2) in which previously acquired original sentences and annotation target phrases included in the previously acquired original sentences are stored in association with each other. According to the translation device 113 of the third modification of the first embodiment of the present disclosure, it is possible to extract annotated phrases that are likely to require the addition of annotation information by referring to phrases that have been extracted in the past as annotated phrases. Furthermore, according to the translation device 113 of the third modification of the first embodiment of the present disclosure, it is possible to add annotations to annotated phrases by referring to the annotation information that has been added to phrases that have been extracted in the past as annotated phrases. Therefore, even if the original text contains phrases that are difficult to translate literally, it is possible to add annotations to the annotated phrases by referring to the past history and display a translation that is easy to understand.

[0108] 8, the past annotated phrases stored in the history information DB2 may be registered in the user dictionary DB1. When the same annotated phrase is stored multiple times in the history information DB2, the annotated phrase may be registered in the user dictionary DB1.

[0109] Variation 4. In the first embodiment and the first to third variations of the first embodiment, we have described translation devices 110-113 that display on a display device annotation information corresponding to annotation target phrases output by annotation target extraction units 20-23. In the fourth variation of the first embodiment, we will describe translation device 114 that accepts selection of annotation target phrases, of which annotation information is to be displayed on a display device, from among the annotation target phrases output by annotation target extraction unit 23.

[0110] A translation device 114 according to the fourth modification of the first embodiment will be described with reference to Fig. 9. Fig. 9 is a functional block diagram showing a translation system 1014 including the translation device 114.

[0111] As shown in FIG. 9, translation device 114 according to Variation 4 of Embodiment 1 differs from translation devices 110-113 according to Embodiment 1 and Variations 1-3 of Embodiment 1 in that translation device 114 further includes a display phrase selection unit 70.

[0112] The display phrase selection unit 70 accepts a selection of annotated phrases, of which annotation information is to be displayed on the display device 400, from among the annotated phrases output by the annotation target extraction unit 23. The display phrase selection unit 70 outputs the accepted selection of annotated phrases to the annotation adding unit 31 and the display processing unit 50. The annotation adding unit 31 adds annotation information only to the annotation target phrases selected by the display phrase selecting unit 70. The display processing unit 50 displays the annotation target phrase selected by the display phrase selection unit 70 in an emphasized manner.

[0113] For example, suppose that the original text is "We have received a call from Kanda GL in Hokushi," and the annotation target extraction unit 23 outputs "Hokushi" and "GL" as annotation target phrases. In this case, the display phrase selection unit 70 accepts a selection of an annotation target phrase for which annotation information is to be displayed on the display device from among the two annotation target phrases output by the annotation target extraction unit 23. When the display phrase selection unit 70 accepts the selection of "Hokushi," the annotation addition unit 31 acquires and adds the annotation target phrase "Hokkaido branch office" corresponding to "Hokushi."

[0114] 5, the display phrase selection unit 70 accepts a selection of an annotation target phrase, of which annotation information is to be displayed on the display device 400, from among the annotation target phrases output by the annotation target extraction unit 23. Then, instead of ST5, the annotation adding unit 31 checks whether the annotation information of the annotation target phrase accepted by the annotation target extraction unit 23 is stored in the user dictionary DB1 or history information DB2.

[0115] The following describes the operations and effects of translation device 114, translation system 1014, and translation method according to Variation 4 of Embodiment 1 of the present disclosure. The translation device 114 according to the fourth variant of the first embodiment of the present disclosure further includes a display phrase selection unit 70 that accepts the selection of an annotation target phrase, from among the annotation target phrases output by the annotation target extraction unit 20-23, for which annotation information is to be displayed on the display device 400. According to the translation device 114 according to the fourth modification of the first embodiment of the present disclosure, it is possible to select phrases for which annotation information needs to be displayed from among the phrases to be annotated, thereby making it possible to display annotation information more effectively for the user.

[0116] Variation 5. In the first embodiment and the third variation of the first embodiment, we have described translation devices 110 and 114 equipped with annotation adding units 30 and 31 that acquire and add annotation information corresponding to words and phrases to be annotated that are stored in the user dictionary DB1 or the history information DB2. In the fifth variation of the first embodiment, we will describe translation device 115 equipped with annotation adding unit 32 that adds annotation information input by a user.

[0117] A translation device 115 according to a fifth modification of the first embodiment will be described with reference to Fig. 10. Fig. 10 is a functional block diagram showing a translation system 1015 including the translation device 115.

[0118] As shown in FIG. 10, translation device 115 according to variant 5 of embodiment 1 differs from translation devices 110-114 according to embodiment 1 and variant 1-4 of embodiment 1 in that translation device 115 further includes an annotation input receiving unit 80 and an annotation adding unit 32 that differs from those in embodiment 1 and variant 1-4 of embodiment 1.

[0119] The annotation input accepting unit 80 accepts input of annotation information corresponding to the annotated phrase from the user via the operation terminal. Specifically, the annotation input accepting unit 80 causes the operation terminal to display an image for accepting input of annotation information corresponding to the annotated phrase. The user inputs annotation information corresponding to the annotated phrase in accordance with the display on the operation terminal. The annotation input accepting unit 80 acquires the annotation information accepted as input at the operation terminal. The annotation adding unit 32 adds annotation information input by the annotation input receiving unit 80 to the annotation target phrase.

[0120] For example, assume that the annotation target extraction unit 23 outputs "GL" as a phrase to be annotated. The annotation input receiving unit 80 receives an input of "group leader" as annotation information corresponding to the phrase to be annotated "GL." The annotation adding unit 31 adds the annotation information "group leader" received by the annotation input receiving unit 80 to the phrase to be annotated "GL" included in the original text.

[0121] The annotation addition unit 31 may replace the annotation addition unit 30 according to embodiment 1 and the annotation addition unit 31 according to variant example 3 of embodiment 1, or may acquire annotation information input into the history information DB2 in addition to the functions of at least one of the annotation addition unit 30 according to embodiment 1 and the annotation addition unit 31 according to variant example 3 of embodiment 1.

[0122] In a translation method using the translation device 115 according to the fifth variant of the first embodiment, instead of or after ST5 shown in FIG. 5, the annotation input accepting unit 80 accepts input of annotation information, and the annotation adding unit 32 adds the annotation information accepted by the annotation input accepting unit 80 to the phrase to be annotated.

[0123] The following describes the operations and effects of translation device 115, translation system 1015, and translation method according to Variation 5 of Embodiment 1 of the present disclosure. The translation device 115 according to the fifth modification of the first embodiment of the present disclosure further includes an annotation input accepting unit 80 that accepts input of annotation information corresponding to the annotated phrase. The annotation adding unit 32 included in the translation device 115 adds the annotation information input accepted by the annotation input accepting unit 80 to the annotated phrase. According to the translation device 115 according to the fifth modification of the first embodiment of the present disclosure, annotation information can be added to the annotated phrase even when the storage unit (user dictionary DB1 or history information DB2) does not store appropriate annotation information corresponding to the annotated phrase.

[0124] Variation 6. In the third variation of the first embodiment, a description has been given of translation device 113 including annotation adding unit 31 that acquires and adds annotation information corresponding to annotated phrases stored in user dictionary DB1 or history information DB2. In the sixth variation of the first embodiment, a description will be given of translation device 116 that accepts a selection of annotation information to be displayed on display device 400 from a plurality of annotation information stored in user dictionary DB1 or history information DB2.

[0125] A translation device according to a sixth modification of the first embodiment will be described with reference to Fig. 11. Fig. 11 is a functional block diagram showing a translation system 1016 including the translation device .

[0126] As shown in FIG. 11, translation device 116 according to variant 6 of embodiment 1 differs from translation devices 1101-115 according to embodiment 1 and variants 1-5 of embodiment 1 in that translation device 116 further includes an annotation selection receiving unit 81 and an annotation adding unit 33 that differs from those in embodiment 1 and variants 1-5 of embodiment 1.

[0127] The user dictionary DB1 or the history information DB2 stores a plurality of annotation information items corresponding to one annotation target phrase. For example, the annotation information for the annotated phrase "Hokushi" includes the definition of the annotated phrase "Hokkaido Branch," the pronunciation of the annotated phrase "Hokushi," and the translation of the annotated phrase into another language, English, "Hokkaido Branch."

[0128] The annotation selection receiving unit 81 receives a selection of annotation information to be displayed on the display device 400 from a plurality of pieces of annotation information stored in the user dictionary DB1 or the history information DB2 from the user via the operation terminal. Specifically, the annotation selection receiving unit 81 causes the operation terminal to display an image for receiving the selection of annotation information to be displayed on the display device 400. The user selects the annotation information to be displayed on the display device 400 according to the display on the operation terminal. The annotation input receiving unit 80 acquires the annotation information selected by the operation terminal. The annotation adding unit 33 adds the annotation information selected by the annotation selection receiving unit 81 to the annotation target phrase.

[0129] For example, the annotation selection receiving unit 81 receives a selection of the definition of the annotated phrase "Hokkaido branch office" from among multiple pieces of annotation information for the annotated phrase "Hokushi" as annotation information to be displayed on the display device 400. The annotation adding unit 33 adds the annotation information "Hokkaido branch office" selected and received by the annotation selection receiving unit 81 to the annotated phrase "Hokushi" included in the original text.

[0130] In addition to performing at least one of the functions of the annotation addition unit 30 according to embodiment 1, the annotation addition unit 31 according to variant 3 of embodiment 1, and the annotation addition unit 32 according to variant 5 of embodiment 1, the annotation addition unit 33 adds annotation information that has been selected.

[0131] In the translation method using the translation device 116 according to the sixth variant of the first embodiment, after ST5 shown in FIG. 5, the annotation selection receiving unit 81 receives a selection of annotation information, and the annotation adding unit 33 adds the annotation information selected and received by the annotation selection receiving unit 81 to the phrase to be annotated.

[0132] The actions and effects of the translation device 116, the translation system 1016, and the translation method according to the sixth modification of the first embodiment of the present disclosure will be described. Translation device 116 according to Variation 6 of Embodiment 1 of the present disclosure further includes an annotation selection receiving unit 81 that receives a selection of annotation information to be displayed on display device 400 from a plurality of pieces of annotation information stored in a storage unit (user dictionary DB1 or history information DB2). An annotation adding unit 33 included in translation device 116 adds the annotation information selected and received by annotation selection receiving unit 81 to the annotated phrase. According to the translation device 116 according to the sixth modification of the first embodiment of the present disclosure, annotation information required by the user can be selected, and therefore annotation information that helps the user to understand can be added with high accuracy.

[0133] Variation 7. In the seventh modification of the first embodiment, a translation device 117 that acquires and adds annotation information using the artificial intelligence unit 200 will be described.

[0134] A translation device 117 according to the seventh modification of the first embodiment will be described with reference to Fig. 12. Fig. 12 is a functional block diagram showing a translation system 1017 including the translation device 117.

[0135] As shown in FIG. 12, translation device 117 according to variant 7 of embodiment 1 differs from translation devices 110-116 according to embodiment 1 and variants 1-6 of embodiment 1 in that it includes annotation addition unit 34 that acquires annotation information using artificial intelligence unit 200 and generates annotated original text.

[0136] The annotation addition unit 34 inputs the annotated phrases obtained by the annotation target extraction unit 23 to the artificial intelligence unit 200, and obtains and outputs an annotated original text from the artificial intelligence unit in which annotation information corresponding to the annotated phrases has been added to the annotated phrases.

[0137] In the translation method using the translation device 117 according to the seventh modification of the first embodiment, instead of steps ST5 and ST6 shown in FIG. 5, the annotation adding unit 34 uses the artificial intelligence unit 200 to generate annotated original text.

[0138] The following describes the operations and effects of the translation device 117, the translation system 1017, and the translation method according to the seventh modification of the first embodiment of the present disclosure. The annotation addition unit 34 provided in the translation device 117 according to the seventh variant of the first embodiment of the present disclosure inputs the annotated phrases acquired by the annotation target extraction unit 20-23 to the artificial intelligence unit 200, and acquires and outputs an annotated original text from the artificial intelligence unit 200 in which annotation information corresponding to the annotated phrases has been added to the annotated phrases. According to the translation device 117 according to the seventh modification of the first embodiment of the present disclosure, an annotated original text can be generated without the user having to register or input annotation information. Therefore, it is possible to reduce the burden on the user and display an easily understandable translation even when the original text contains words that are difficult to translate literally.

[0139] Embodiment 2 In the first embodiment of the present disclosure, a translation device 110 that adds annotation information to annotated phrases and performs translation has been described. In the second embodiment of the present disclosure, a translation device 120 that adds annotation information corresponding to attributes of listeners to annotated phrases will be described.

[0140] First, a translation system 1020 including a translation device 120 according to the second embodiment of the present disclosure will be described with reference to Fig. 13. Fig. 13 is a functional block diagram of the translation system 1020 according to the second embodiment.

[0141] As shown in FIG. 13, translation device 120 according to embodiment 2 differs from translation device 110 according to embodiment 1 in that it further includes an attribute receiving unit 90 and an attribute information DB3, and in that it includes an annotation adding unit 35 that is different from that of embodiment 1.

[0142] The attribute receiving unit 90 receives input of the attributes of the audience from the user via the operation terminal. Specifically, the attribute receiving unit 90 displays the input of the attributes of the audience on the operation terminal. The user inputs the attributes of the audience according to the display on the operation terminal. The phrase designation receiving unit 60 acquires the attributes of the audience that have been input via the operation terminal. Here, the attributes of the listeners include the listeners' age, hometown, or place of stay.

[0143] The annotation adding unit 35 acquires annotation information corresponding to the attribute accepted by the attribute accepting unit 90. Specifically, the annotation adding unit 35 inputs the attribute accepted by the attribute accepting unit 90 to the artificial intelligence unit 200, and refers to the attribute information DB3 to acquire from the artificial intelligence unit 200 annotation information converted to correspond to the attribute, and outputs the annotation information.

[0144] The attribute information DB3 stores conversion rules for annotation information corresponding to attributes. For example, the attribute information DB3 stores characters, words, or phrases to be learned according to the age of the listener, or languages ​​or dialects according to the listener's hometown and place of residence.

[0145] The annotation addition unit 35 inputs the annotation target words and phrases obtained by the annotation target extraction unit 20, the attributes received by the attribute receiving unit 90, and a description instructing the annotation information to be converted to correspond to the attributes by referring to the attribute information DB3 to the artificial intelligence unit 200, and obtains the annotation information converted to correspond to the attributes from the artificial intelligence unit 200 as the annotation target words and phrases.

[0146] The annotation adding unit 35 need only be able to acquire annotation information corresponding to the attributes accepted by the attribute accepting unit 90 without using the artificial intelligence unit 200, and does not necessarily have to use the artificial intelligence unit 200. For example, the user may register annotation information corresponding to the attributes in the user dictionary DB1 in advance, and the annotation adding unit 35 may acquire the annotation information corresponding to the attributes accepted by the attribute accepting unit 90 from the user dictionary DB1 in which annotation information corresponding to the attributes is stored.

[0147] In addition to performing at least one of the functions of the annotation addition unit 30 according to embodiment 1, the annotation addition unit 31 according to variant 3 of embodiment 1, the annotation addition unit 32 according to variant 5 of embodiment 1, the annotation addition unit 33 according to variant 6 of embodiment 1, and the annotation addition unit 34 according to variant 7 of embodiment 1, the annotation addition unit 35 acquires annotation information corresponding to attributes.

[0148] In the translation method using the translation device 120 according to the second embodiment, instead of ST5 shown in FIG. 5, or before or after ST5, the annotation adding unit 35 acquires annotation information corresponding to the attribute received by the attribute receiving unit 90.

[0149] The following describes the operations and effects of translation device 120, translation system 1020, and translation method according to the second embodiment of the present disclosure. The translation device 120 according to the second embodiment of the present disclosure further includes an attribute receiving unit 90 that receives input of a user's attributes. The annotation adding unit 35 included in the translation device 120 acquires annotation information corresponding to the attributes received by the attribute receiving unit 90. The translation device 120 according to the second embodiment of the present disclosure can add annotation information corresponding to the attributes of the listeners to the annotation target phrases. Therefore, annotation information that helps the listeners understand can be added with high accuracy.

[0150] Embodiment 3 In the second embodiment of the present disclosure, a translation device 120 is described that adds annotation information corresponding to the attributes of the audience to the annotated phrase. In the third embodiment of the present disclosure, a translation device 130 is described that further acquires annotation-related information related to the annotated phrase from the artificial intelligence unit 200 and further displays it on the display device 400.

[0151] First, a translation system 1030 including a translation device 130 according to the third embodiment of the present disclosure will be described with reference to Fig. 14. Fig. 14 is a functional block diagram of the translation system 1030 according to the third embodiment.

[0152] As shown in FIG. 14, translation device 130 according to embodiment 3 differs from translation devices 110-120 according to embodiment 1-2 in that it further includes a related information acquisition unit 91 and a related information DB4, and in that it includes a display processing unit 51 that is different from that of embodiment 1-2.

[0153] The related information acquisition unit 91 inputs the annotated phrase output by the annotation target extraction unit 20 to the artificial intelligence unit 200, and acquires annotation-related information related to the annotated phrase from the artificial intelligence unit 200. Here, the annotation-related information refers to a photograph, illustration, video, or document indicating the annotated phrase, or a URL (Uniform Resource Locator) indicating the location of such information. For example, if the annotated phrase is a phrase indicating a work location, the related information acquisition unit 91 acquires a photograph, map, or URL of explanatory material regarding the work location.

[0154] Specifically, the related information acquisition unit 91 refers to the related information DB4 and acquires annotation related information related to the annotated phrase from the artificial intelligence unit 200. The related information DB4 stores annotation related information registered in advance by the user.

[0155] The related information acquisition unit 91 may cause the artificial intelligence unit 200 to refer to an internal company database or the like, in addition to the related information database 4. For example, if the phrase to be annotated is an employee name, the related information acquisition unit 91 acquires a photo of the employee extracted by the artificial intelligence unit 200 from the internal company database.

[0156] The annotation-related information may be output by processing the original information extracted by the artificial intelligence unit 200. For example, if the phrase to be annotated is an employee's name, the artificial intelligence unit 200 may extract a photo in which the person appears from the internal database, and the related information acquisition unit 91 may acquire, as annotation-related information, the photo in which only the periphery of the person is cropped.

[0157] In addition, the related information acquisition unit 91 may input a description to the artificial intelligence unit 200 instructing it to generate annotation-related information related to the annotated phrase and the annotation-related information related to the annotated phrase, and acquire the annotation-related information from the artificial intelligence unit 200.

[0158] The related information acquisition unit 91 outputs the annotation related information acquired from the artificial intelligence unit 200 to the display processing unit 51 . The display processing unit 51 further displays the annotation related information acquired by the related information acquisition unit 91 .

[0159] In the translation method using the translation device 130 according to embodiment 3, simultaneously with or after ST8 shown in FIG. 5, the related information acquisition unit 91 acquires annotation-related information related to the phrase to be annotated from the artificial intelligence unit 200, and the display processing unit 51 further displays the annotation-related information acquired by the related information acquisition unit 91.

[0160] The following describes the operations and effects of the translation device 130, the translation system 1030, and the translation method according to the third embodiment of the present disclosure. The translation device 130 according to the third embodiment of the present disclosure further includes a related information acquisition unit 91 that inputs the annotation target phrases output by the annotation target extraction units 20-23 to the artificial intelligence unit 200 and acquires and outputs annotation related information related to the annotation target phrases from the artificial intelligence unit 200. The display processing unit 51 included in the translation device 130 further displays the annotation related information acquired by the related information acquisition unit 91. According to the translation device 130 according to the third embodiment of the present disclosure, annotation-related information indicating the annotated phrase can be displayed together with annotation information corresponding to the annotated phrase. Therefore, a translation of a sentence that includes a phrase that is difficult to translate literally can be displayed in a manner that is easier for the user to understand.

[0161] Embodiment 4 In the first embodiment of the present disclosure, a translation device 110 is described that translates an annotated original text in which annotation information is added to annotated phrases. In the fourth embodiment of the present disclosure, a translation device 140 is described that translates the annotated original text into multiple languages ​​after back-translating the annotated original text into a language different from the original text.

[0162] First, a translation system 1040 including a translation device 140 according to the fourth embodiment of the present disclosure will be described with reference to Fig. 15. Fig. 15 is a functional block diagram of the translation system 1040 according to the fourth embodiment.

[0163] As shown in FIG. 15, the translation device 140 according to the fourth embodiment further includes a back translation unit 41, a correction receiving unit 42, and a second translation unit 43, and differs from the translation devices 110-130 according to the first to third embodiments in that it includes a display processing unit 52 that is different from that of the first to third embodiments.

[0164] The back translation unit 41 generates an annotated back-translated text by back-translating the annotated translation by the translation unit 40. Here, back translation means translating a translation that has been translated into a language different from the original text back into the original language. Specifically, the back translation unit 41 provides the annotated translation to the translation engine 300 and obtains the annotated back translation from the translation engine 300 . The back translation unit 41 outputs the annotated back translation to the display processing unit 52.

[0165] The correction receiving unit 42 receives corrections to the annotated original text after the display processing unit 52 displays the annotated back-translated text. The correction receiving unit 42 outputs the received corrected annotated original text to the translation unit 40 and the second translation unit 43.

[0166] When the correction receiving unit 42 receives a correction to the annotated original text, the second translation unit 43 generates an annotated second translation by translating the corrected annotated original text received by the correction receiving unit 42 into a second translation target language. The second translation target language is a language different from the original text. When the correction receiving unit 42 does not receive a correction to the annotated original text, the second translation unit 43 generates an annotated second translation by translating the annotated original text generated by the annotation adding unit 30 into the second translation target language. Specifically, the second translation unit 43 provides the translation engine 300 with the annotated original text after correction received by the correction receiving unit 42 or the annotated original text generated by the annotation adding unit 30, and obtains the annotated second translation text from the translation engine 300. The second translation unit 43 outputs the annotated second translated sentence to the display processing unit 52.

[0167] In addition to the original text, words to be annotated, annotation information, annotated original text, and annotated translation displayed by the display processing unit 50 according to embodiment 1, the display processing unit 52 also displays the annotated back-translation generated by the back translation unit 41 and the annotated second translation generated by the second translation unit 43. The display processing unit may also switch the display to a mode in which only the annotated back-translation and the annotated second translation are displayed by switching tabs or the like.

[0168] Next, a translation method using translation device 140 according to embodiment 4 will be described with reference to Fig. 16. Fig. 16 is a flowchart showing the translation method according to embodiment 4. Note that steps ST3 to ST7 shown in Fig. 5 are omitted in Fig. 16.

[0169] In the translation method using the translation device 140 according to the fourth embodiment, first, steps ST1 to ST8 are performed in the same manner as in the first embodiment. For the sake of convenience, it is assumed here that the original Japanese text "GL Kanda was concerned about the status of attendance at the training course." is acquired, "GL" is extracted as the phrase to be annotated, and the annotated original text "GL Kanda (group leader) was concerned about the status of attendance at the training course." to which "group leader" is added as annotation information, and the annotated translation into which the annotated original text is translated are displayed on the display device 400.

[0170] After step ST8, the back translation unit 41 acquires the annotated translation from the translation unit 40 and generates an annotated back translation by back-translating the annotated translation by the translation unit 40 (step ST11). The back translation unit 41 outputs the back translation to the display processing unit 52.

[0171] The display processing unit 52 displays the back-translated sentence output by the back translation unit 41 on the display device 400 (step ST12).

[0172] The user looks at the back-translated text displayed on the display device 400 and considers whether to revise the annotated original text. For example, when the above annotated translation is back-translated, it is expected that the back-translated text displayed on the display device 400 will be "GL Kanda (group leader) was concerned about the status of his own training attendance." If the nuance that the user wants to convey in the annotated original text is "GL Kanda (group leader) was concerned about the status of employees' training attendance," the user will need to revise the annotated original text and translate it again.

[0173] The correction receiving unit 42 receives corrections to the annotated original text after the annotated back-translated text is displayed by the display processing unit 52. When the user performs an operation to correct the annotated original text (YES in step ST13), the correction receiving unit 42 receives the correction to the annotated original text (step ST14). The corrected annotated original text is input to the artificial intelligence unit 200, which again extracts words and phrases to be annotated (step ST2). Here, for the sake of convenience, it is assumed that the correction receiving unit 42 receives the correction "Kanda GL (group leader) was concerned about the status of the training attendance of the people from North China."

[0174] When the correction receiving unit 42 receives a correction to the annotated original text (step ST14), the correction receiving unit 42 outputs the corrected annotated original text to the annotation target extraction unit 20. The annotation target extraction unit 20 inputs the corrected annotated original text to the artificial intelligence unit 200, and extracts annotation target phrases again (steps ST2 to ST4). When new annotation target phrases are output by the annotation target extraction unit 20, the annotation adding unit 30 generates an annotated original text by adding annotation information to the newly output annotation target phrases, and outputs the annotated original text to the translation unit 40 (steps ST5 to ST6). For example, if the annotated original text after correction is "Kanda GL (Group Leader) was concerned about the status of the training attendance of the people in Hokushi," the annotation target extraction unit 20 extracts "Hokushi," which was added during correction, as a new phrase to be annotated, and outputs it to the annotation addition unit 30. The annotation addition unit 30 adds annotation information "Hokkaido Branch" corresponding to the newly output annotation target phrase "Hokushi."

[0175] When the annotation adding unit 30 generates and outputs a new annotated original text (step ST6), the translation unit 40 and the back translation unit 41 translate the original text and display it on the display device 400 (steps ST7 to ST12). The above steps are repeated until the correction accepting unit 42 no longer accepts corrections.

[0176] When the user does not perform an operation to correct the annotated original text (NO in step ST13) and the correction receiving unit 42 stops accepting corrections to the annotated original text, the second translation unit 43 acquires the annotated original text that was last corrected by the correction receiving unit 42 or the annotated original text generated by the annotation adding unit 30. The second translation unit 43 generates an annotated second translation by translating any of the acquired annotated original texts into the second translation target language (step ST15). Specifically, the second translation unit 43 provides any of the acquired annotated original texts to the translation engine 300 and acquires the annotated second translation from the translation engine 300. The second translation unit 43 outputs the annotated second translation to the display processing unit 52.

[0177] The display processing unit 52 causes the display device 400 to display the annotated second translated sentence output by the second translation unit 43 (step ST16).

[0178] The operation and effects of the translation device 140 and the translation method according to the fourth embodiment of the present disclosure will be described. Translation device 140 according to embodiment 4 of the present disclosure further includes a back translation unit 41 that generates an annotated back-translated sentence by translating an annotated translation by translation unit 40 into an original language, a correction receiving unit 42 that receives corrections to the annotated original sentence, and a second translation unit 43 that, when a correction to the annotated original sentence is received by correction receiving unit 42, generates an annotated second translation by translating the annotated original sentence after the correction received by correction receiving unit 42 into a second translation target language. Display processing unit 52 included in translation device 140 further displays the annotated back-translated sentence generated by back translation unit 41 and the annotated second translation generated by second translation unit 43.

[0179] According to the translation device 140 and the translation method according to the fourth embodiment of the present disclosure, a sentence containing words that are difficult to translate literally can be annotated and then translated into multiple languages, and the time required for translating into multiple languages ​​can be reduced. The time required for translation can be reduced because the translation into the second target language can be performed with content that reflects the translation results into the target language. For example, when the original text is written in a user's native language, the original text may contain a mixture of colloquial and written expressions, resulting in a sentence that is understandable to a native speaker but grammatically ambiguous. For this reason, it is generally considered difficult to translate the original text into a correct translation in one translation run, and the original text must be corrected. The translation device 140 translates the annotated original text into the target language, performs back-translation to display the annotated back-translated text, and then the user checks the annotated back-translated text and corrects the original text. The corrected annotated original text is then used to translate into the second target language. Therefore, translation into the second target language can be performed using the corrected annotated original text, which is already easy to translate, thereby achieving a translation with higher translation accuracy than conventional methods in a shorter time. Furthermore, even when the similarity between the grammar in the language used in the original text and the grammar in the second language used in the translation target is greater than the similarity between the grammar in the language used in the original text and the grammar in the language used in the translation target, translation device 140 translates into the second language used in the translation target using the original text that the user has checked and corrected through back translation, thereby making it possible to obtain a translation with higher translation accuracy in a shorter time than before. Furthermore, when a phrase that is difficult to translate literally is added or deleted during correction, the annotation information added to the phrase to be annotated can be changed, thereby preventing oversight of adding annotation information.

[0180] Although the above describes an example in which translation device 110-140 is provided separately from display device 400, translation device 110-140 and display device 400 may be provided as an integrated unit. For example, translation device 110-140 may be provided as an integrated function of an operation terminal included in display device 400, and the translation result may be displayed on an audience terminal included in display device 400. Furthermore, translation device 110-140, the operation terminal, and the audience terminal may be provided as an integrated unit.

[0181] Although the present disclosure has been described above based on the embodiments, the present disclosure is not limited to the embodiments. Furthermore, appropriate combinations, modifications, omissions, etc. of the embodiments are also included within the scope of the technical idea of ​​the present disclosure. [Explanation of symbols]

[0182] 1 User Dictionary DB 2. History Information DB 3 Attribute information DB 4 Related Information Database 10 Original text acquisition department 20, 21, 22, 23 Annotation target extraction part 30, 31, 32, 33, 34, 35, 36, 37 Annotation section 40 Translation Department 41 Back Translation Department 42 Correction Reception Department 43 Second Translation Department 50, 51, 52 Display processing unit 60 Word specification reception section 70 Display phrase selection section 80 Annotation input reception section 81 Annotation selection reception section 90 Attribute reception section 91 Related Information Acquisition Department 110, 111, 112, 113, 114, 115, 116, 117, 120, 130, 140 Translation devices 200 Artificial Intelligence Department 201 Trained model memory unit 202 Model control section 300 Translation Engine 400 display device 1010, 1011, 1012, 1013, 1014, 1015, 1016, 1017, 1020, 1030, 1040 Translation Systems

Claims

1. an original text acquisition unit that acquires an original text to be translated; an annotation target extraction unit that inputs the original text acquired by the original text acquisition unit into an artificial intelligence unit, and acquires and outputs annotation target words and phrases that are difficult to translate and are included in the original text from the artificial intelligence unit; an annotation adding unit that acquires annotation information corresponding to the annotation target phrases output by the annotation target extraction unit, and generates annotated original text by adding the annotation information to the annotation target phrases; a translation unit that generates an annotated translation by translating the annotated original text generated by the annotation adding unit into a translation target language; a display processing unit that displays the annotated translation translated by the translation unit on a display device; A translation device comprising:

2. The annotation target extraction unit causes the artificial intelligence unit to translate the original text into a translation target language, and acquires and outputs from the artificial intelligence unit, as the annotation target phrases, phrases that are difficult to translate by the artificial intelligence unit. The translation device according to claim 1 .

3. a phrase specification receiving unit that receives specification of the phrase to be annotated, The annotation target extraction unit acquiring from the artificial intelligence unit the annotation target phrases included in the original text and corresponding to the specification accepted by the phrase specification accepting unit, and outputting the annotation target phrases; The translation device according to claim 1 .

4. The annotation target extraction unit The artificial intelligence unit refers to a storage unit in which the words to be annotated that have been registered in advance by a user are stored, and the artificial intelligence unit acquires and outputs the words to be annotated that are included in the original text and are stored in the storage unit. The translation device according to claim 1 .

5. The annotation target extraction unit making the artificial intelligence unit refer to a storage unit in which the original sentence obtained in the past and the phrases to be annotated contained in the original sentence obtained in the past are stored in association with each other, and obtaining and outputting the phrases to be annotated contained in the original sentence and stored in the storage unit from the artificial intelligence unit; The translation device according to claim 1 .

6. The annotation adding unit acquiring the annotation information corresponding to the annotation target phrase output by the annotation target extraction unit from a storage unit in which the annotation target phrase and the annotation information registered in advance by a user are stored in association with each other; The translation device according to claim 1 .

7. The storage unit The annotation information stores at least one of a definition of the annotated phrase, a reading of the annotated phrase, and an expression obtained by translating the definition of the annotated phrase into another language.

5. The translation device according to claim 4.

8. the storage unit stores a plurality of pieces of annotation information corresponding to the annotation target phrases; an annotation selection receiving unit that receives a selection of the annotation information to be displayed on the display device from the plurality of annotation information stored in the storage unit; the annotation adding unit adds the annotation information selected by the annotation selection receiving unit to the annotation target phrase; 5. The translation device according to claim 4.

9. the annotation adding unit adds at least one of a definition of the annotated phrase, a pronunciation of the annotated phrase, and an expression obtained by translating the definition of the annotated phrase into another language to the annotated phrase as the annotation information; The translation device according to claim 1 .

10. the translation unit translates the annotated phrase using any one of a definition of the annotated phrase, a reading of the annotated phrase, and an expression obtained by translating the definition of the annotated phrase into another language; The translation device according to claim 1 .

11. a display phrase selection unit that accepts a selection of an annotation target phrase for which the annotation information is to be displayed on the display device, from among the annotation target phrases output by the annotation target extraction unit; The translation device according to claim 1 .

12. further comprising an annotation input receiving unit that receives input of the annotation information corresponding to the annotation target phrase; The annotation adding unit adding the annotation information inputted and accepted by the annotation input accepting unit to the annotation target phrase; The translation device according to claim 1 .

13. The annotation adding unit inputs the annotation target phrases acquired by the annotation target extraction unit to the artificial intelligence unit, and acquires and outputs the annotated original text in which the annotation information corresponding to the annotation target phrases is added to the annotation target phrases from the artificial intelligence unit. The translation device according to claim 1 .

14. further comprising an attribute receiving unit that receives input of a user's attribute; the annotation adding unit acquires the annotation information corresponding to the attribute accepted by the attribute accepting unit. The translation device according to claim 1 .

15. The display processing unit further displaying the annotated source text so as to highlight the phrases to be annotated; The translation device according to claim 1 .

16. a related information acquisition unit that inputs the annotation target phrase output by the annotation target extraction unit to the artificial intelligence unit, and acquires annotation related information related to the annotation target phrase from the artificial intelligence unit and outputs the information; the display processing unit further displays the annotation related information acquired by the related information acquisition unit. The translation device according to claim 1 .

17. a back translation unit that generates an annotated back-translation by translating the annotated translation by the translation unit into the language of the original text; a correction receiving unit that receives corrections to the annotated original text; a second translation unit that, when the correction receiving unit receives a correction to the annotated original text, generates a second annotated translation by translating the annotated original text after the correction received by the correction receiving unit into a second translation target language; Furthermore, the display processing unit further displays the annotated back-translated sentence generated by the back translation unit and the annotated second translated sentence generated by the second translation unit. The translation device according to claim 1 .

18. The annotated phrase is an abbreviation, a buzzword, a technical term, an in-house term, or an industry term. The translation device according to claim 1 .

19. A translation device according to any one of claims 1 to 18; an artificial intelligence unit that, when the original text is input from the translation device, outputs the annotation target phrases included in the original text to the translation device; A translation system comprising:

20. A computer comprising: An original text acquisition step of acquiring an original text to be translated; an annotation target extraction step of inputting the original text acquired in the original text acquisition step into an artificial intelligence unit, and acquiring and outputting annotation target words and phrases that are difficult to translate and are included in the original text from the artificial intelligence unit; an annotation adding step of acquiring annotation information corresponding to the annotation target phrases output in the annotation target extracting step, and generating an annotated original text by adding the annotation information to the original text; a translation step of generating an annotated translation by translating the annotated original text generated in the annotation adding step into a translation target language; a display control step of displaying the annotated translation obtained in the translation step on a display device; A translation method that performs:

Images