Text translation method, device, equipment and storage medium

Abstract

The embodiments of the present application provide a text translation method, apparatus, device and storage medium, and relate to the field of text translation technology. The method obtains an initial text and a translation element, inputs the translation element into a grammatical parsing agent, a rhetorical understanding agent and a cultural background agent to obtain grammatical parsing data, rhetorical parsing data and cultural parsing data, inputs the initial text, grammatical parsing data, rhetorical parsing data and cultural parsing data into a task allocation agent for task allocation to obtain a first ratio of grammatical parsing data, a second ratio of rhetorical parsing data and a third ratio of cultural parsing data, generates initial parsing data based on the first ratio, the second ratio and the third ratio, inputs the initial text and the initial parsing data into a translation agent for translation processing, and obtains a translated text of the initial text. Combined with the collective wisdom of multiple agents, information such as rhetoric and cultural differences is comprehensively utilized in the translation process to improve the accuracy of the translation results.

Description

Technical Field

[0001] The present application relates to the field of text translation technology, and in particular to a text translation method, apparatus, device, and storage medium. Background Art

[0002] Machine translation technology, a key application in natural language processing, aims to automatically convert one language into another through computer programs. With the advancement of globalization and the increase in cross-cultural communication, machine translation is playing an increasingly important role in promoting international communication and information acquisition.

[0003] Neural network models are usually used in related technologies to implement the translation process. However, related neural network models usually rely on a single network architecture and lack multi-perspective and multi-level semantic understanding capabilities. As a result, the current translation results are insufficient in semantic coherence, contextual adaptability and cultural communication, making it difficult to meet the standards of "faithfulness, expressiveness and elegance". Summary of the Invention

[0004] The main purpose of the embodiments of the present application is to propose a text translation method, apparatus, device and storage medium to improve the accuracy of translation results.

[0005] To achieve the above objectives, a first aspect of an embodiment of the present application provides a text translation method, comprising:

[0006] Obtaining an initial text and obtaining a translation element corresponding to the initial text;

[0007] Input the translation elements into the grammar parsing agent, rhetoric understanding agent and cultural background agent respectively for multi-angle analysis to obtain corresponding grammar parsing data, rhetoric parsing data and cultural parsing data;

[0008] Inputting the initial text, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data into a task assignment agent for task assignment, thereby obtaining a first ratio of the grammatical parsing data, a second ratio of the rhetorical parsing data, and a third ratio of the cultural parsing data;

[0009] generating initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data;

[0010] The initial text and the initial parsed data are input into a text translation agent for translation processing to obtain a translated text of the initial text.

[0011] In some embodiments, before generating initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data, the method further includes:

[0012] Inputting the initial text, the first ratio, the second ratio and the third ratio into an interactive evaluation agent for data evaluation to generate evaluation data;

[0013] If the evaluation data indicates that at least one of the first ratio, the second ratio and the third ratio needs to be adjusted, the evaluation data is sent to the task assignment agent, and the task assignment agent is used to regenerate the first ratio, the second ratio and the third ratio based on the evaluation data, the initial text, the grammatical parsing data, the rhetorical parsing data and the cultural parsing data until the evaluation data indicates that no adjustment is required.

[0014] In some embodiments, generating initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data includes:

[0015] Acquiring first feature information of the grammatical analysis data, acquiring second feature information of the rhetorical analysis data, and acquiring third feature information of the cultural analysis data;

[0016] The first ratio is multiplied by the first feature information, the second ratio is multiplied by the second feature information, and the third ratio is multiplied by the third feature information, and the multiplication results are added to obtain the initial parsed data.

[0017] In some embodiments, inputting the initial text and the initial parsed data into a text translation agent for translation processing to obtain a translated text of the initial text includes:

[0018] Optimize and adjust the translation process based on the initial text and the initial parsed data to obtain translation reference data;

[0019] The initial text and the translation reference data are input into the text translation agent for translation processing to obtain the translated text.

[0020] In some embodiments, optimizing and adjusting the translation process based on the initial text and the initial parsed data to obtain translation reference data includes:

[0021] Taking the initial text and the initial parsed data as input data;

[0022] Inputting the input data into a grammar translation agent for grammar translation to obtain grammar translation data;

[0023] Inputting the input data into a semantic translation agent for semantic translation to obtain semantic translation data;

[0024] Inputting the input data into a style translation agent for style translation to obtain style translation data;

[0025] The translation reference data is obtained according to the grammatical translation data, the semantic translation data and the style translation data.

[0026] In some embodiments, inputting the initial text and the translation reference data into the text translation agent for translation processing to obtain the translated text includes:

[0027] Inputting the initial text and the translation reference data into the text translation agent for translation processing to obtain comprehensive translation data;

[0028] The comprehensive translation data is optimized and adjusted to obtain the translation text.

[0029] In some embodiments, optimizing and adjusting the translation of the comprehensive translation data to obtain the translation text includes:

[0030] Inputting the comprehensive translation data into a fluency optimization agent to adjust the fluency to obtain fluency optimization data;

[0031] Inputting the comprehensive translation data into a style optimization agent for style optimization to obtain style translation data;

[0032] Inputting the comprehensive translation data into a consistency optimization agent for consistency optimization to obtain consistency optimization data;

[0033] The fluency optimization data, the style translation data and the consistency optimization data are input into a data coordination intelligent body for information integration to obtain the translation text.

[0034] In some embodiments, the method further comprises:

[0035] Acquiring a target document format, and adjusting the format of the translated text based on the target document format to obtain a formatted output text;

[0036] Performing a grammar check and / or a spelling check on the format output text to obtain a final translated text.

[0037] In some embodiments, obtaining the translation element corresponding to the initial text includes:

[0038] Segmenting the initial text to obtain at least one text segmentation;

[0039] Performing part-of-speech tagging on the text segmentation to obtain part-of-speech data corresponding to each text segmentation;

[0040] Performing syntactic analysis on the initial text to obtain syntactic component data;

[0041] The translation element is obtained based on the text segmentation, the part-of-speech data and the syntactic component data.

[0042] To achieve the above-mentioned purpose, a second aspect of the embodiments of the present application provides a text translation device, comprising:

[0043] Data acquisition module: used to acquire the initial text and obtain the translation elements corresponding to the initial text;

[0044] Multi-angle parsing module: used to input the translation elements into the grammar parsing agent, rhetoric understanding agent and cultural background agent respectively for multi-angle parsing, and obtain corresponding grammar parsing data, rhetoric parsing data and cultural parsing data;

[0045] A task assignment module is configured to input the initial text, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data into a task assignment agent for task assignment, thereby obtaining a first ratio of the grammatical parsing data, a second ratio of the rhetorical parsing data, and a third ratio of the cultural parsing data;

[0046] An initial data generation module: configured to generate initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data;

[0047] Translation processing module: used for inputting the initial text and the initial parsed data into the text translation agent for translation processing to obtain a translated text of the initial text.

[0048] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0049] To achieve the above-mentioned purpose, the fourth aspect of the embodiment of the present application proposes a storage medium, which is a storage medium. The storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0050] The text translation method, apparatus, device and storage medium proposed in the embodiments of the present application obtain an initial text and translation elements corresponding to the initial text, input the translation elements into a grammatical parsing agent, a rhetorical understanding agent and a cultural background agent for multi-angle analysis, and obtain corresponding grammatical parsing data, rhetorical parsing data and cultural parsing data. The initial text, grammatical parsing data, rhetorical parsing data and cultural parsing data are input into a task allocation agent for task allocation, and a first proportion of the grammatical parsing data, a second proportion of the rhetorical parsing data and a third proportion of the cultural parsing data are obtained. Initial parsing data is generated according to the first proportion, the second proportion, the third proportion, the grammatical parsing data, the rhetorical parsing data and the cultural parsing data. Finally, the initial text and the initial parsing data are input into a translation agent for translation processing to obtain a translated text of the initial text. In the embodiment of the present application, different translation perspectives such as grammar, rhetoric and cultural background are analyzed respectively, and then the proportion of analysis results of different translation perspectives is determined based on the initial text through task allocation of intelligent agents, thereby obtaining initial analysis data, and adjusting the translation results of the initial text based on the initial analysis data. Combined with the collective wisdom of multiple intelligent agents, information such as rhetoric and cultural differences is comprehensively utilized in the translation process to achieve a faithful, expressive and elegant translation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a flowchart of the text translation method provided in an embodiment of the present application.

[0052] Figure 2 This is an overall flow chart of the text translation method provided in the embodiment of the present application.

[0053] Figure 3 This is a flowchart of obtaining translation elements corresponding to the initial text provided by an embodiment of the present application.

[0054] Figure 4 This is a flowchart of adjusting the first ratio, the second ratio, and the third ratio provided in an embodiment of the present application.

[0055] Figure 5 This is a flowchart of generating initial parsing data based on the first ratio, the second ratio, the third ratio, grammatical parsing data, rhetorical parsing data and cultural parsing data provided in an embodiment of the present application.

[0056] Figure 6 This is a flowchart of an embodiment of the present application for optimizing and adjusting the translation process based on the initial text and initial parsed data to obtain translation reference data.

[0057] Figure 7 This is a flowchart of optimizing and adjusting the translation of comprehensive translation data to obtain a translated text, provided in an embodiment of the present application.

[0058] Figure 8 This is a structural block diagram of a text translation device provided in another embodiment of the present application.

[0059] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0061] It should be noted that although the functional modules are divided in the device schematic and the logical order is shown in the flowchart, in some cases, the steps shown or described can be performed in a different order than the module division in the device or the order in the flowchart.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0063] First, let’s analyze some of the terms used in this application:

[0064] Artificial intelligence (AI) is a new technical discipline that studies and develops theories, methods, technologies, and application systems for simulating, extending, and expanding human intelligence. A branch of computer science, AI seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thinking. It also encompasses the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0065] Machine translation technology, a key application in natural language processing, aims to automatically convert one language into another through computer programs. With the advancement of globalization and the increase in cross-cultural communication, machine translation is playing an increasingly important role in promoting international communication and information acquisition.

[0066] Related technologies often employ neural network models to implement translation. However, these models typically rely on a single network architecture and lack multi-perspective, multi-layered semantic understanding capabilities. This results in current translations lacking in semantic coherence, contextual adaptability, and cultural communication, making it difficult to achieve the standards of "faithfulness, expressiveness, and elegance." "Faithfulness, expressiveness, and elegance" means that translations must not only be faithful to the original meaning (faithfulness), but also be fluent (expressiveness) and possess literary beauty (elegance). Therefore, faithful, expressive, and elegant translations require translating more complex language, metaphors, and cultural nuances across multiple languages. To compensate for the limitations of single neural network models, related technologies include multi-model integration techniques to improve translation accuracy and diversity, effectively reducing common translation errors and improving the naturalness and fluency of translation results to a certain extent. However, these technologies still lack multi-perspective semantic understanding capabilities, making it difficult to generate natural and accurate translations when faced with complex translation tasks, such as complex language structures, long sentences, and sentences with complex grammar. Due to insufficient understanding of information such as rhetoric and cultural differences, it is difficult for the translation results to accurately convey the subtle meanings corresponding to these specific cultural backgrounds and contexts. The specific expressions of culture, idioms and contexts cannot be accurately expressed, and the translation may appear stiff or unnatural in the target language culture.

[0067] Based on this, the embodiments of the present application provide a text translation method, apparatus, device and storage medium, which analyze different translation angles such as grammar, rhetoric and cultural background respectively, and then determine the proportion of analysis results of different translation angles based on the initial text through task allocation of intelligent agents, thereby obtaining initial analysis data, adjusting the translation results of the initial text based on the initial analysis data, and utilizing the collective wisdom of multiple intelligent agents based on the interaction process between multiple intelligent agents. In the translation process, information such as rhetoric and cultural differences is comprehensively utilized to achieve a faithful, expressive and elegant translation effect.

[0068] The embodiments of the present application provide a text translation method, apparatus, device, and storage medium, which are specifically illustrated by the following embodiments. First, the text translation method in the embodiments of the present application is described.

[0069] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that machines have the functions of perception, reasoning and decision-making.

[0070] Artificial intelligence (AI) technology is a comprehensive discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Foundational AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0071] The text translation method provided by the embodiment of the present application relates to the field of text translation technology. The text translation method provided by the embodiment of the present application can be applied to a terminal, can also be applied to a server side, and can also be a computer program running in a terminal or a server side. For example, a computer program can be a native program or software module in an operating system; it can be a local (Native) application (Application, APP), that is, a program that needs to be installed in the operating system to run, such as a client that supports text translation, that is, a program that can be run only by downloading it to a browser environment; it can also be a small program that can be embedded in any APP. In short, the above-mentioned computer program can be an application, module or plug-in in any form. Wherein, the terminal communicates with the server via a network. The text translation method can be executed by a terminal or a server, or performed in collaboration by a terminal and a server.

[0072] In some embodiments, the terminal can be a smart phone, tablet computer, laptop computer, desktop computer or smart watch, etc. In addition, the terminal can also be an intelligent vehicle-mounted device. The intelligent vehicle-mounted device applies the text translation method of this embodiment to provide related services to enhance the driving experience. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms; it can also be a service node in a blockchain system, in which each service node in the blockchain system forms a peer-to-peer (P2P) network. The P2P protocol is an application layer protocol running on top of the Transmission Control Protocol (TCP). The terminal and the server can be connected via Bluetooth, Universal Serial Bus (USB) or a network, and this embodiment does not limit this.

[0073] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0074] The text translation method in the embodiment of the present application is described below.

[0075] Figure 1 This is an optional flowchart of the text translation method provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps 110 to 150. It is also understood that this embodiment is Figure 1 The order of steps 110 to 150 is not specifically limited, and the order of steps can be adjusted or some steps can be reduced or added according to actual needs.

[0076] Step 110: Obtain the initial text and obtain the translation element corresponding to the initial text.

[0077] In one embodiment, the initial text is the source text that needs to be translated, which needs to be translated into a target language. The target language and the language of the source text may be the same or different. When the two are the same, the translation process can be understood as a process of parsing the initial text, which is not limited in this embodiment.

[0078] Reference Figure 2 , Figure 2 This is an overall flow chart of the text translation method provided by the embodiment of this application. Figure 2 After obtaining the initial text, preprocessing is required to obtain translation elements for participating in the subsequent translation process to improve the accuracy of the translation results.

[0079] In one embodiment, referring to Figure 3 , Figure 3 This is a flowchart of obtaining translation elements corresponding to the initial text provided by an embodiment of the present application, which specifically includes the following steps:

[0080] Step 310: Segment the initial text to obtain at least one text segment.

[0081] In one embodiment, the purpose of segmenting the initial text to obtain text segmentation is to more accurately identify each word in the initial text through the segmentation process, avoiding misinterpretation of polysemous words or phrases as single words. This helps to more accurately understand the meaning of the initial text during the subsequent translation process.

[0082] Step 320: Perform part-of-speech tagging on the text segmentation to obtain part-of-speech data corresponding to each text segmentation.

[0083] In one embodiment, part-of-speech tagging is performed on each obtained text segmentation, thereby obtaining part-of-speech data corresponding to each text segmentation. Part-of-speech tagging helps to understand the role of the text segmentation in the entire original text.

[0084] Step 330: Perform syntactic analysis on the initial text to obtain syntactic component data.

[0085] In one embodiment, the purpose of syntactic analysis is to generate a syntax tree of the original text as syntactic component data. The syntax tree can represent the overall structure of the original text, which helps to accurately grasp the context and semantic relationship of the original text during the translation process, thereby generating a more accurate translation.

[0086] Step 340: Obtain translation elements based on the text segmentation, part-of-speech data, and syntactic component data.

[0087] In one embodiment, text segmentation, part-of-speech data, and syntactic component data are collectively used as translation elements. The translation elements contain the sentence structure of the original text. During the translation process, vocabulary that is more in line with the expression habits of the target language can be more clearly selected to construct the translation of the original text, and the word order and sentence structure can be adjusted according to the grammatical rules of the target language. While maintaining the accuracy of the translation, the style and tone of the original text can be retained as much as possible by analyzing features such as part of speech and sentence structure, making the translation more natural and fluent.

[0088] Step 120: Input the translation elements into the grammar parsing agent, rhetoric understanding agent and cultural background agent respectively for multi-angle analysis to obtain corresponding grammar parsing data, rhetoric parsing data and cultural parsing data.

[0089] In one embodiment, referring to Figure 2 After obtaining the translation elements, the translation elements are used as input information for the grammar parsing agent, the rhetoric understanding agent and the cultural background agent respectively. The grammar parsing agent is used to perform preliminary analysis of the grammar to obtain grammar parsing data, the rhetoric understanding agent is used to perform preliminary analysis of the rhetoric to obtain rhetoric parsing data, and the cultural background agent is used to perform preliminary analysis of the cultural background to obtain cultural parsing data.

[0090] In one embodiment, a grammatical parsing agent analyzes the grammatical structure of the original text based on translation elements, more accurately grasping the relationship between different text segmentations, understanding the true meaning of the original text, and avoiding mistranslations caused by misunderstandings. Especially for original texts containing multiple clauses or modifiers, only by clarifying the grammatical relationship between each text segmentation can the original message of the original text be accurately conveyed. Therefore, embodiments of the present application utilize grammatical parsing data to store the grammatical information obtained through preliminary parsing.

[0091] In one embodiment, the rhetoric-related information of the initial text is analyzed based on the translation elements through a rhetoric understanding agent. Rhetoric, as an important means of language expression, carries rich emotional colors and cultural heritage. The embodiment of the present application uses a rhetoric understanding agent to carefully analyze the rhetorical techniques in the initial text to obtain rhetoric parsing data, so as to more accurately grasp the meaning and emotional tone of the initial text. Among them, rhetorical techniques can be metaphors, personification, symbols, metaphors, etc. In the subsequent translation process, expressions corresponding to the rhetoric in the initial text are found based on the target language to improve the accuracy and fluency of the translation results. It can be understood that the rhetoric understanding agent is trained based on a large number of figurative expressions, symbolic expressions, metaphorical expressions and other corpora to ensure the accuracy and cultural adaptability of the rhetoric parsing process.

[0092] In one embodiment, the cultural background of the initial text includes deep information such as history, tradition, custom, and values ​​in the source language, and this information is often reflected in the form of metaphors, allusions, idioms, etc. in the language. Therefore, the embodiment of the present application utilizes the cultural background intelligent agent to analyze these cultural elements and obtain cultural analysis data. In the subsequent translation process, the cultural analysis data is used to more accurately understand the meaning and context of the initial text and avoid misunderstandings caused by literal translation. At the same time, the cultural analysis data can also be used to make appropriate cultural adaptation and conversion in the subsequent translation results, so that the translation results are closer to the cultural cognition of the target language, thereby improving the accuracy and readability of the translation results.

[0093] It is understandable that the above-mentioned different intelligent agents all have corresponding prompt words when inputting, which are used to guide the corresponding intelligent agents to perform target parsing operations, and this embodiment does not limit this.

[0094] Step 130: Input the initial text, grammatical parsing data, rhetorical parsing data and cultural parsing data into the task assignment agent for task assignment, and obtain a first proportion of grammatical parsing data, a second proportion of rhetorical parsing data and a third proportion of cultural parsing data.

[0095] In one embodiment, even if the grammatical parsing data, rhetorical parsing data and cultural parsing data corresponding to the initial text are obtained through the above-mentioned grammatical parsing agent, rhetorical understanding agent and cultural background agent, these data are still in an independent state, and it is necessary to organically combine these data to fully utilize the collective intelligence among the grammatical parsing agent, rhetorical understanding agent and cultural background agent to produce the effect of 1+1>2.

[0096] Therefore, refer to Figure 2 The embodiment of the present application then introduces a task assignment agent to implement the interaction process between the grammatical parsing agent, the rhetorical understanding agent, and the cultural background agent. The initial text, grammatical parsing data, rhetorical parsing data, and cultural parsing data are input into the task assignment agent for task assignment, resulting in a first ratio of grammatical parsing data, a second ratio of rhetorical parsing data, and a third ratio of cultural parsing data. The purpose of inputting the initial text here is twofold: one is to avoid the gradual forgetting of the initial task during the translation process, and the other is to select the degree of participation of the grammatical parsing data, rhetorical parsing data, and cultural parsing data in the translation result based on the overall meaning of the initial text, and quantify the degree of participation in the form of the first ratio, the second ratio, and the third ratio.

[0097] The following are some examples to illustrate this.

[0098] For example, if the initial text is a legal document, the translation process requires extremely high precision and rigorous grammatical structure to ensure accurate interpretation of the terms. Therefore, the first proportion of grammatical analysis data corresponds to the highest proportion, for example, 60%. Furthermore, legal documents generally avoid the use of decorative language and pursue objective and direct expression. Therefore, the second proportion of rhetorical analysis data corresponds to the lowest proportion, for example, 5%. Finally, while legal terminology is universal, certain legal concepts may be interpreted differently in different jurisdictions, requiring understanding of the relevant cultural background. Therefore, the third proportion of cultural analysis data corresponds to a moderate proportion, for example, 35%.

[0099] For example, if the initial text is a literary work, while grammatical correctness is important during the translation process, adjustments to the grammatical structure may sometimes be made to maintain the style and rhythm of the original work. Therefore, the first percentage of grammatical analysis data corresponds to a moderate proportion, such as 30%. Furthermore, the charm of literary works largely comes from their rhetorical techniques, such as metaphor and personification, which require careful processing to preserve the artistic effect of the original text. Therefore, the second percentage of rhetorical analysis data corresponds to a higher proportion, such as 30%. Finally, literary works are full of references and metaphors from specific cultures. Understanding and conveying these cultural connotations is crucial to the translation results. Therefore, the third percentage of cultural analysis data corresponds to the highest proportion, such as 40%.

[0100] For example, the initial text is a technical manual or instruction manual. During the translation process, technical documents require clear and accurate instructions. Grammatical errors may lead to operational errors or misunderstandings. Therefore, the first proportion of grammatical analysis data corresponds to the highest proportion, for example, 85%. In addition, technical documents focus on the direct transmission of information and need to avoid unnecessary rhetorical embellishments. Therefore, the second proportion of rhetorical analysis data corresponds to the lowest proportion, for example, 0%. Finally, even if the general understanding of users in different regions must be considered, technical terms and operating procedures are generally universal and not easily affected by cultural differences. Therefore, the third proportion of cultural analysis data corresponds to a moderate proportion, for example, 15%.

[0101] From the above, we can see that when assigning tasks, the task assignment agent will intelligently adjust the importance of grammar, rhetoric and cultural background in the translation results based on the different types of initial texts received.

[0102] In one embodiment, since the first ratio, the second ratio and the third ratio directly affect the final translation result, the embodiment of the present application also provides an evaluation measure to further improve the distribution accuracy of these three ratios. Figure 4 , Figure 4 This is a flowchart of adjusting the first ratio, the second ratio, and the third ratio provided in an embodiment of the present application, which specifically includes the following steps:

[0103] Step 410: Input the initial text, the first ratio, the second ratio and the third ratio into the interactive evaluation agent for data evaluation to generate evaluation data.

[0104] In one embodiment, an interactive evaluation agent is introduced to evaluate the first, second, and third ratios based on the inputs, generating evaluation data that indicates whether these ratios need to be adjusted. It is understood that the evaluation data can be an evaluation indicator that reflects whether the first, second, and third ratios need to be adjusted.

[0105] Step 420: If the evaluation data indicates that at least one of the first ratio, the second ratio and the third ratio needs to be adjusted, the evaluation data is sent to the task assignment agent, and the task assignment agent is used to regenerate the first ratio, the second ratio and the third ratio based on the evaluation data, the initial text, the grammatical parsing data, the rhetorical parsing data and the cultural parsing data until the evaluation data indicates that no adjustment is required.

[0106] In one embodiment, referring to Figure 2 If the evaluation data determines that at least one of the first, second, and third ratios needs to be adjusted, the interactive evaluation agent will send the evaluation data to the task assignment agent. After receiving the evaluation data, the task assignment agent will reassign tasks based on the initial text, grammatical parsing data, rhetorical parsing data, and cultural parsing data, and regenerate new first, second, and third ratios. These will be sent to the interactive evaluation agent for evaluation. After multiple evaluations, if the interactive evaluation agent believes that these ratio values ​​are reasonable, it will generate evaluation data that does not require adjustment to the first, second, and third ratios. It is understandable that such ratio adjustments may not be perfected in one go and may require multiple inquiries and generation between the interactive evaluation agent and the task assignment agent.

[0107] Step 140: Generate initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data.

[0108] In one embodiment, after obtaining the first ratio of the grammatical parsing data, the second ratio of the rhetorical parsing data, and the third ratio of the cultural parsing data through the above process, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data are fused according to the first ratio, the second ratio, and the third ratio to generate initial parsed data.

[0109] In one embodiment, referring to Figure 5 , Figure 5This is a flowchart of generating initial parsed data based on the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data provided by an embodiment of the present application, specifically including the following steps:

[0110] Step 510: Acquire first feature information of the grammatical parsing data, acquire second feature information of the rhetorical parsing data, and acquire third feature information of the cultural parsing data.

[0111] In one embodiment, feature extraction is performed by word embedding or sentence embedding to obtain first feature information of grammatical parsing data, second feature information of rhetorical parsing data, and third feature information of cultural parsing data. These feature data are all represented as feature vectors.

[0112] Step 520: multiply the first ratio by the first feature information, multiply the second ratio by the second feature information, and multiply the third ratio by the third feature information, and add the multiplication results to obtain initial parsed data.

[0113] In one embodiment, the feature vectors obtained according to the first ratio, the second ratio and the third ratio are weighted and averaged. Specifically, the first ratio is multiplied by the first feature information, the second ratio is multiplied by the second feature information, and the third ratio is multiplied by the third feature information. The multiplication results are added together to obtain a weighted average vector, and the vector is converted into text to obtain initial parsed data.

[0114] Step 150: Input the initial text and initial parsed data into the text translation agent for translation processing to obtain a translated text of the initial text.

[0115] In one embodiment, after the initial parsed data is obtained, it and the initial text are translated by inputting a text translation agent. The purpose of inputting the initial text is to allow the text translation agent not to forget the initial starting point of the translation task, thereby improving the accuracy of the translated text.

[0116] In one embodiment, in order to further improve the accuracy of the translated text, the initial text and initial parsed data need to be processed accordingly. Specifically, the translation process is optimized and adjusted based on the initial text and initial parsed data to obtain translation reference data, and then the initial text and translation reference data are input into the text translation agent for translation processing to obtain the translated text.

[0117] In one embodiment, referring to Figure 6 , Figure 6 This is a flowchart of optimizing and adjusting the translation process based on the initial text and initial parsed data to obtain translation reference data, provided by an embodiment of the present application, specifically including the following steps:

[0118] Step 610: Take the initial text and initial parsed data as input data.

[0119] In one embodiment, referring to Figure 2 , the optimization and adjustment process needs to be carried out based on the initial text and the initial parsed data as input.

[0120] Step 620: Input the input data into the grammar translation agent for grammar translation to obtain grammar translation data.

[0121] Step 630: Input the input data into the semantic translation agent for semantic translation to obtain semantic translation data.

[0122] Step 640: Input the input data into the style translation agent for style translation to obtain style translation data.

[0123] In one embodiment, the initial parsed data contains data information related to grammar, rhetoric, and cultural background. At this time, the grammar translation agent, semantic translation agent, and style translation agent are used in parallel to perform translation from corresponding angles to focus on different aspects of the initial text.

[0124] Grammatical translation primarily focuses on sentence structure, specifically grammatical rules. This translation method emphasizes maintaining the syntactic structure of the original text, converting the language word by word and sentence by sentence without much consideration of context or overall meaning. Therefore, the present embodiment utilizes a grammatical translation agent to perform a grammatically focused translation of the initial text based on the initial parsed data, and uses the resulting grammatically relevant translation data as grammatical translation data.

[0125] Semantic translation, on the other hand, focuses more on the meaning of the text, that is, the accuracy of the content. During the semantic translation process, the original text's intention and information will be expressed as much as possible, rather than just its literal meaning. This requires a deep understanding of the original text and the ability to find the most appropriate expression in the target language to convey the same meaning. Therefore, the embodiment of this application uses a semantic translation agent to perform a semantic-focused translation of the original text based on the initial parsed data, and the resulting semantically relevant translation data is used as the semantic translation data.

[0126] Stylistic translation, on the other hand, focuses on the stylistic characteristics of a text, including the author's personality, stylistic characteristics, and the text's emotional overtones. This type of translation not only aims to convey the content and meaning of the original text, but also to reproduce the style and tone of the original text as closely as possible. Therefore, this embodiment of the application utilizes a style translation agent to perform a style-focused translation of the initial text based on the initial parsed data, and the resulting style-related translation data is used as the style translation data.

[0127] For example, a sentence in the initial text: "The quick brown fox jumps over the lazy dog" may be translated into "The quick brown fox jumps over the lazy dog" after grammatical translation. The grammatical translation data retains the syntactic structure of the original text. The semantic translation data obtained through semantic translation focuses more on conveying meaning and may be translated into: "A quick brown fox jumps over a lazy dog". The style translation data obtained through style translation may further consider the style of the initial text based on the rhetoric or cultural background related information in the initial parsed data. For example, if the initial text is a humorous story, it may be translated into: "The quick brown fox jumped over the lazy dog ​​lightly". It will be understood that only one sentence is used as an example here, and the embodiments of the present application will also consider the content association of the context in the initial text during the translation process.

[0128] Step 650: Obtain translation reference data according to the grammatical translation data, the semantic translation data, and the style translation data.

[0129] In one embodiment, the grammatical translation data is the content after grammatical translation, the semantic translation data is the content after semantic alignment, and the style translation data is the content after style adaptation. Therefore, the translation reference data can be obtained by summarizing these three aspects of content.

[0130] Next, refer to Figure 2 , inputting the initial text and translation reference data into the text translation intelligent agent for translation processing to obtain the translated text specifically includes: inputting the initial text and translation reference data into the text translation intelligent agent for translation processing to obtain comprehensive translation data, optimizing and adjusting the translation of the comprehensive translation data to obtain the translated text.

[0131] In one embodiment, after obtaining translation reference data related to grammar, semantics and style, the text translation agent uses it as reference information in the translation process of the initial text to obtain comprehensive translation data, which can be verified using multi-party information to improve the overall translation accuracy.

[0132] In one embodiment, after the comprehensive translation data is obtained, it can be output as the final translation text. However, the comprehensive translation data can be further optimized and adjusted to improve the output data quality. Figure 7 , Figure 7 This is a flowchart of optimizing and adjusting the translation of comprehensive translation data to obtain a translation text provided by an embodiment of the present application, which specifically includes the following steps:

[0133] Step 710: Input the comprehensive translation data into the fluency optimization agent to perform fluency adjustment to obtain fluency optimization data.

[0134] In one embodiment, the comprehensive translation data is input into a fluency optimization agent, which performs fluency adjustments on the text corresponding to the comprehensive translation data, such as adjusting sentence structure, vocabulary selection, and grammatical correction, to ensure that the text corresponding to the fluency optimization data after adjustment reads more naturally and conforms to the language habits of the target language.

[0135] Step 720: Input the comprehensive translation data into the style optimization agent to perform style optimization to obtain style translation data.

[0136] In one embodiment, the comprehensive translation data is input into a style optimization agent, which focuses on the text style corresponding to the comprehensive translation data to ensure that the translation process not only conveys the original meaning of the original text, but also retains the stylistic features of the original text, such as formality, informality, humor, etc. The style of the text corresponding to the obtained consistency optimization data is consistent with the style of the original text.

[0137] Step 730: Input the comprehensive translation data into the consistency optimization agent to perform consistency optimization to obtain consistency optimization data.

[0138] In one embodiment, the comprehensive translation data is input into a consistency optimization agent, which ensures that the translation process maintains consistency in terms of terminology, tense, person, etc., and avoids inconsistencies in the same content in the same text.

[0139] Step 740: Input the fluency optimization data, style translation data, and consistency optimization data into the data coordination agent for information integration to obtain a translated text.

[0140] In one embodiment, the fluency optimization data, style translation data, and consistency optimization data generated in parallel are input into a data coordination agent for information integration to produce a translated text. The data coordination agent unifies the three optimization data sets through the information integration process, ensuring that the resulting translated text is consistent and free of conflict or duplication.

[0141] In one embodiment, referring to Figure 2 After obtaining the translated text, post-processing operations can be performed. Specifically, this includes obtaining the target document format, such as font type, size, color, paragraph arrangement, heading hierarchy, headers and footers, and chart layout. The translated text is then formatted based on the target document format, resulting in output text that is consistent with the target document format. Next, the formatted output text can be grammatically checked and / or spelled using natural language processing tools to improve the readability and accuracy of the translation results, ultimately obtaining and outputting the final translated text.

[0142] Refer to the following Figure 2, which generally describes the execution process of the text translation method of the embodiment of the present application.

[0143] Figure 2 The process can be divided into four stages, the first of which is the preliminary parsing stage. Initial text is obtained, followed by preprocessing. The initial text is segmented to obtain at least one text segmentation, and then part-of-speech data is tagged to obtain the corresponding part-of-speech data for each text segmentation. The initial text is then syntactically analyzed to obtain syntactic component data. Based on the text segmentation, part-of-speech data, and syntactic component data, the translation elements corresponding to the initial text are generated. The translation elements are then fed into a grammatical parsing agent, a rhetorical understanding agent, and a cultural background agent for multi-angle analysis, obtaining the corresponding grammatical parsing data, rhetorical parsing data, and cultural parsing data.

[0144] Then, the second stage begins, which is the task assignment stage. The initial text, grammatical parsing data, rhetorical parsing data, and cultural parsing data are input into the task assignment agent for task assignment, resulting in a first ratio for the grammatical parsing data, a second ratio for the rhetorical parsing data, and a third ratio for the cultural parsing data. The initial text, first ratio, second ratio, and third ratio are then input into the interactive evaluation agent for data evaluation, generating evaluation data. If the evaluation data indicates that at least one of the first, second, and third ratios needs to be adjusted, the evaluation data is sent to the task assignment agent, which then regenerates the first, second, and third ratios based on the evaluation data, the initial text, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data until the evaluation data indicates that no adjustment is required. Initial parsing data is then generated based on the final first ratio, second ratio, third ratio, grammatical parsing data, rhetorical parsing data, and cultural parsing data.

[0145] Next, we enter the third stage, which is the multi-angle translation stage. The initial text and initial parsed data are used as input data. The input data is fed into the grammatical translation agent for grammatical translation to obtain grammatical translation data. The input data is fed into the semantic translation agent for semantic translation to obtain semantic translation data. The input data is fed into the style translation agent for style translation to obtain style translation data. Based on the grammatical translation data, semantic translation data, and style translation data, translation reference data is obtained.

[0146] The final stage is the output optimization stage. First, the initial text and translation reference data are input into the text translation agent for translation processing to obtain comprehensive translation data. Next, the comprehensive translation data is optimized and adjusted to obtain the translated text. Specifically, the comprehensive translation data is input into the fluency optimization agent for fluency adjustment to obtain fluency optimization data; the comprehensive translation data is input into the style optimization agent for style optimization to obtain style translation data; the comprehensive translation data is input into the consistency optimization agent for consistency optimization to obtain consistency optimization data; the fluency optimization data, style translation data, and consistency optimization data are input into the data coordination agent for information integration to obtain the translated text. Finally, adjustments can be made to the output format to obtain the target document format, such as font type, size, color, paragraph arrangement, title hierarchy, headers and footers, and chart layout. The translated text is then formatted based on the target document format to obtain a formatted output text consistent with the target document format. The formatted output text can also be grammatically checked and / or spelled using natural language processing tools to improve the readability and accuracy of the translation results, and the final translated text is obtained and output.

[0147] It can be seen that multi-level optimization of the translation text through the above process can ensure that the final translation text is not only accurate but also conforms to the usage habits and cultural background of the target language.

[0148] It is understandable that the agent in the embodiment of the present application is implemented based on a large language model. Each agent is trained in advance according to the corresponding sample, has suitable prompt words and model parameters, and can achieve the specific functions such as required depth semantic understanding, grammatical conversion, and cultural background adaptation. The architecture design of this multi-agent collaboration utilizes multiple different agents to interact and share information, combines the task allocation process to achieve effective collaboration and mutual division of labor in the translation process, realizes the improvement of group intelligence, can analyze and understand the initial text from multiple perspectives and levels, significantly improves the quality of translation, and ensures that the translation result reaches high standards in terms of grammar, semantics, and cultural adaptability. Especially in complex translation tasks, it performs well, for example, in translation tasks such as processing long sentences, complex grammar, and cross-cultural content, in the embodiment of the present application, multi-agent can provide highly natural and fluent translation results by collaborating.

[0149] The technical solution provided by the embodiment of the present application obtains an initial text and the translation elements corresponding to the initial text, and inputs the translation elements into a grammar parsing agent, a rhetoric understanding agent, and a cultural background agent for multi-angle analysis, thereby obtaining corresponding grammar parsing data, rhetoric parsing data, and cultural parsing data. The initial text, grammar parsing data, rhetoric parsing data, and cultural parsing data are input into a task allocation agent for task allocation, thereby obtaining a first ratio of grammar parsing data, a second ratio of rhetoric parsing data, and a third ratio of cultural parsing data. Initial parsing data is generated based on the first ratio, the second ratio, the third ratio, the grammar parsing data, the rhetoric parsing data, and the cultural parsing data. Finally, the initial text and the initial parsing data are input into a translation agent for translation processing, thereby obtaining a translated text of the initial text. In the embodiment of the present application, different translation angles such as grammar, rhetoric, and cultural background are parsed respectively, and then the proportion of parsing results of different translation angles is determined based on the initial text through the task allocation of the agent, thereby obtaining initial parsing data. The translation result of the initial text is adjusted based on the initial parsing data. In combination with the collective wisdom of multiple agents, information such as rhetoric and cultural differences is comprehensively utilized in the translation process to achieve a faithful and elegant translation effect.

[0150] The present application also provides a text translation device that can implement the above text translation method. Figure 8 , the device comprises:

[0151] The data acquisition module 810 is used to acquire the initial text and the translation elements corresponding to the initial text.

[0152] Multi-angle parsing module 820: used to input translation elements into the grammar parsing agent, rhetoric understanding agent and cultural background agent respectively for multi-angle parsing, and obtain corresponding grammar parsing data, rhetoric parsing data and cultural parsing data.

[0153] Task assignment module 830: used to input the initial text, grammatical parsing data, rhetorical parsing data and cultural parsing data into the task assignment agent for task assignment, and obtain a first proportion of grammatical parsing data, a second proportion of rhetorical parsing data and a third proportion of cultural parsing data.

[0154] Initial data generation module 840: used to generate initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data and the cultural parsing data.

[0155] The translation processing module 850 is used to input the initial text and the initial parsed data into the text translation agent for translation processing to obtain a translated text of the initial text.

[0156] The specific implementation of the text translation device of this embodiment is basically the same as the specific implementation of the above-mentioned text translation method, and will not be repeated here.

[0157] An embodiment of the present application further provides an electronic device, including:

[0158] at least one memory;

[0159] at least one processor;

[0160] at least one program;

[0161] The program is stored in the memory, and the processor executes the at least one program to implement the text translation method described above. The electronic device can be any smart terminal including a mobile phone, a tablet computer, a personal digital assistant (PDA), an in-vehicle computer, etc.

[0162] See also Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is shown. The electronic device includes:

[0163] The processor 901 may be implemented as a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is configured to execute relevant programs to implement the technical solutions provided in the embodiments of the present application.

[0164] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 902 and is called by the processor 901 to execute the text translation method of the embodiments of this application.

[0165] Input / output interface 903, used to implement information input and output;

[0166] Communication interface 904, used to implement communication interaction between this device and other devices, which can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.); and

[0167] Bus 905 , which transmits information between various components of the device (e.g., processor 901 , memory 902 , input / output interface 903 , and communication interface 904 );

[0168] The processor 901 , the memory 902 , the input / output interface 903 and the communication interface 904 are connected to each other in communication within the device via a bus 905 .

[0169] An embodiment of the present application further provides a storage medium, which is a storage medium storing a computer program. When the computer program is executed by a processor, the above-mentioned text translation method is implemented.

[0170] The memory, as a non-transient storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0171] The text translation method, apparatus, device and storage medium proposed in the embodiments of the present application obtain an initial text and translation elements corresponding to the initial text, input the translation elements into a grammatical parsing agent, a rhetorical understanding agent and a cultural background agent for multi-angle analysis, and obtain corresponding grammatical parsing data, rhetorical parsing data and cultural parsing data. The initial text, grammatical parsing data, rhetorical parsing data and cultural parsing data are input into a task allocation agent for task allocation, and a first proportion of the grammatical parsing data, a second proportion of the rhetorical parsing data and a third proportion of the cultural parsing data are obtained. Initial parsing data is generated according to the first proportion, the second proportion, the third proportion, the grammatical parsing data, the rhetorical parsing data and the cultural parsing data. Finally, the initial text and the initial parsing data are input into a translation agent for translation processing to obtain a translated text of the initial text. In the embodiment of the present application, different translation perspectives such as grammar, rhetoric and cultural background are analyzed respectively, and then the proportion of analysis results of different translation perspectives is determined based on the initial text through task allocation of intelligent agents, thereby obtaining initial analysis data, and adjusting the translation results of the initial text based on the initial analysis data. Combined with the collective wisdom of multiple intelligent agents, information such as rhetoric and cultural differences is comprehensively utilized in the translation process to achieve a faithful, expressive and elegant translation effect.

[0172] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0173] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0174] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0175] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0176] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0177] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0178] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the above-mentioned units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0179] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0180] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0181] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: various media that can store programs, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0182] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A text translation method, characterized in that: include: Obtaining an initial text and obtaining a translation element corresponding to the initial text; Input the translation elements into the grammar parsing agent, rhetoric understanding agent and cultural background agent respectively for multi-angle analysis to obtain corresponding grammar parsing data, rhetoric parsing data and cultural parsing data; Inputting the initial text, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data into a task assignment agent for task assignment, thereby obtaining a first ratio of the grammatical parsing data, a second ratio of the rhetorical parsing data, and a third ratio of the cultural parsing data; generating initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data; The initial text and the initial parsed data are used as input data, the input data is input into a grammatical translation agent for grammatical translation to obtain grammatical translation data, the input data is input into a semantic translation agent for semantic translation to obtain semantic translation data, the input data is input into a style translation agent for style translation to obtain style translation data, translation reference data is obtained based on the grammatical translation data, the semantic translation data and the style translation data, the initial text and the translation reference data are input into a text translation agent for translation processing to obtain a translated text of the initial text.

2. The text translation method according to claim 1, characterized in that Before generating initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data, the method further includes: Inputting the initial text, the first ratio, the second ratio and the third ratio into an interactive evaluation agent for data evaluation to generate evaluation data; If the evaluation data indicates that at least one of the first ratio, the second ratio and the third ratio needs to be adjusted, the evaluation data is sent to the task assignment agent, and the task assignment agent is used to regenerate the first ratio, the second ratio and the third ratio based on the evaluation data, the initial text, the grammatical parsing data, the rhetorical parsing data and the cultural parsing data until the evaluation data indicates that no adjustment is required.

3. The text translation method according to claim 1, characterized in that Generating initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data includes: Acquiring first feature information of the grammatical analysis data, acquiring second feature information of the rhetorical analysis data, and acquiring third feature information of the cultural analysis data; The first ratio is multiplied by the first feature information, the second ratio is multiplied by the second feature information, and the third ratio is multiplied by the third feature information, and the multiplication results are added to obtain the initial parsed data.

4. The text translation method according to claim 1, characterized in that The step of inputting the initial text and the translation reference data into a text translation agent for translation processing to obtain the translated text includes: Inputting the initial text and the translation reference data into the text translation agent for translation processing to obtain comprehensive translation data; The comprehensive translation data is optimized and adjusted to obtain the translation text.

5. The text translation method according to claim 4, characterized in that: The step of optimizing and adjusting the translation of the comprehensive translation data to obtain the translation text includes: Inputting the comprehensive translation data into a fluency optimization agent to adjust the fluency to obtain fluency optimization data; Inputting the comprehensive translation data into a style optimization agent for style optimization to obtain style translation data; Inputting the comprehensive translation data into a consistency optimization agent for consistency optimization to obtain consistency optimization data; The fluency optimization data, the style translation data and the consistency optimization data are input into a data coordination intelligent body for information integration to obtain the translation text.

6. The text translation method according to any one of claims 1 to 5, characterized in that: The method further comprises: Acquiring a target document format, and adjusting the format of the translated text based on the target document format to obtain a formatted output text; Performing a grammar check and / or a spelling check on the format output text to obtain a final translated text.

7. The text translation method according to any one of claims 1 to 5, characterized in that: The obtaining of the translation element corresponding to the initial text includes: Segmenting the initial text to obtain at least one text segmentation; Performing part-of-speech tagging on the text segmentation to obtain part-of-speech data corresponding to each text segmentation; Performing syntactic analysis on the initial text to obtain syntactic component data; The translation element is obtained based on the text segmentation, the part-of-speech data and the syntactic component data.

8. A text translation device, characterized in that: include: Data acquisition module: used to acquire the initial text and obtain the translation elements corresponding to the initial text; Multi-angle parsing module: used to input the translation elements into the grammar parsing agent, rhetoric understanding agent and cultural background agent respectively for multi-angle parsing, and obtain corresponding grammar parsing data, rhetoric parsing data and cultural parsing data; A task assignment module is configured to input the initial text, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data into a task assignment agent for task assignment, thereby obtaining a first ratio of the grammatical parsing data, a second ratio of the rhetorical parsing data, and a third ratio of the cultural parsing data; An initial data generation module: configured to generate initial parsing data according to the first ratio, the second ratio, the third ratio, the grammatical parsing data, the rhetorical parsing data, and the cultural parsing data; Translation processing module: used to take the initial text and the initial parsed data as input data, input the input data into a grammatical translation agent for grammatical translation to obtain grammatical translation data, input the input data into a semantic translation agent for semantic translation to obtain semantic translation data, input the input data into a style translation agent for style translation to obtain style translation data, obtain translation reference data based on the grammatical translation data, the semantic translation data and the style translation data, and input the initial text and the translation reference data into a text translation agent for translation processing to obtain a translated text of the initial text.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the text translation method according to any one of claims 1 to 7 when executing the computer program.

10. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the text translation method according to any one of claims 1 to 7 is implemented.

Images