Translation quality analysis method, device, equipment, medium and program product

Abstract

The application provides a translation quality analysis method, device, equipment, medium and program product, and relates to the technical field of information processing. The method comprises the following steps: obtaining translation task information, wherein the translation task information comprises at least one of machine translation text, first modification mark information of the machine translation text, post-translation translation text, second modification mark information of the post-translation translation text and standard translation text; determining modification mark quality information of the translation task information according to a first similarity between the machine translation text and the standard translation text; determining post-translation translation quality information of the translation task information according to a second similarity between the post-translation translation text and the standard translation text and the second modification mark information; and determining translation quality information corresponding to the translation task information according to the post-translation translation quality information and the modification mark quality information.

Description

Technical Field

[0001] This invention relates to the field of information processing technology, and in particular to a method, apparatus, device, medium, and program product for analyzing translation quality. Background Technology

[0002] With advancements in society and computer technology, the translation industry is undergoing a series of transformations, with most translation work now being conducted and completed online. DOT (Distributed Online Translation) breaks down large translation tasks into many smaller tasks using specific algorithms and distributes them to multiple translators. This can lead to inconsistent quality in delivered translations, making it difficult to guarantee the overall project quality.

[0003] Therefore, how to analyze the quality of each translation task is an urgent problem that the industry needs to solve. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, medium, and program product for analyzing translation quality, in order to address the shortcomings of existing technologies where the quality of delivered translation tasks varies, thus affecting the overall project delivery quality.

[0005] This invention provides a method for analyzing translation quality, comprising:

[0006] Obtain translation task information, which includes at least one of the following: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and standard translation;

[0007] Based on the first similarity between the machine-translated text and the standard translation, the modification mark quality information of the translation task information is determined;

[0008] Based on the second similarity between the post-translated text and the standard translation, and the second modification marker information, the post-translation quality information of the translation task information is determined;

[0009] The translation quality information corresponding to the translation task information is determined based on the post-translation quality information and the modification mark quality information.

[0010] According to a translation quality analysis method provided by the present invention, based on a first similarity between the machine-translated text and the standard translation, modification marker quality information of the translation task information is determined, including:

[0011] The modification category of the machine-translated text is determined based on the first similarity between the machine-translated text and the standard translation.

[0012] Based on the modification category and the first modification marker information, the modification marker quality information of the translation task information is determined.

[0013] According to a translation quality analysis method provided by the present invention, the post-translation quality information of the translation task information is determined based on the second similarity between the post-translated text and the standard translation, and the second modification marker information, including:

[0014] Based on the second similarity between the translated text and the standard translation, the translation quality information of the translated text is determined;

[0015] Based on the second modification mark information and the translation quality information, a post-translation review evaluation is obtained;

[0016] Based on the translated text quality information and the post-translation review and evaluation, the post-translation quality information of the translation task information is determined.

[0017] According to a translation quality analysis method provided by the present invention, the translation quality information of the translated text is determined based on a second similarity between the translated text and the standard translation, including:

[0018] The second similarity between the translated text and the standard translation is determined based on the edit distance, Jaccard distance, and bilingual evaluation substitution distance between the translated text and the standard translation.

[0019] The translation quality information of the translated text is determined based on the second similarity and the first preset coefficient.

[0020] According to a translation quality analysis method provided by the present invention, a post-translation review evaluation is obtained based on the second modification mark information and the translated text quality information, including:

[0021] When the second modification marker information is the return modification marker information, the post-translation review evaluation is obtained based on the difference between the second preset coefficient and the translation quality information. The return modification marker information includes: minor modification marker information, moderate modification marker information, or severe modification marker information.

[0022] If the second modification marker information is that no modification marker information is required, the translation quality information will be used as the post-translation review and evaluation.

[0023] The present invention also provides a translation quality analysis device, comprising:

[0024] The acquisition module is used to acquire translation task information, which includes at least one of the following: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and standard translation.

[0025] The first determining module is used to determine the modification mark quality information of the translation task information based on the first similarity between the machine-translated text and the standard translation;

[0026] The second determining module is used to determine the post-translation quality information of the translation task information based on the second similarity between the post-translated text and the standard translation and the second modification mark information;

[0027] The third determining module is used to determine the translation quality information corresponding to the translation task information based on the post-translation quality information and the modification mark quality information.

[0028] According to a translation quality analysis apparatus provided by the present invention, based on a first similarity between the machine-translated text and the standard translation, modification marker quality information of the translation task information is determined, including:

[0029] The modification category of the machine-translated text is determined based on the first similarity between the machine-translated text and the standard translation.

[0030] Based on the modification category and the first modification marker information, the modification marker quality information of the translation task information is determined.

[0031] According to a translation quality analysis device provided by the present invention, the post-translation quality information of the translation task information is determined based on the second similarity between the post-translated text and the standard translation, and the second modification marker information, including:

[0032] Based on the second similarity between the translated text and the standard translation, the translation quality information of the translated text is determined;

[0033] Based on the second modification mark information and the translation quality information, a post-translation review evaluation is obtained;

[0034] Based on the translated text quality information and the post-translation review and evaluation, the post-translation quality information of the translation task information is determined.

[0035] According to a translation quality analysis device provided by the present invention, the translation quality information of the translated text is determined based on a second similarity between the translated text and the standard translation, including:

[0036] The second similarity between the translated text and the standard translation is determined based on the edit distance, Jaccard distance, and bilingual evaluation substitution distance between the translated text and the standard translation.

[0037] The translation quality information of the translated text is determined based on the second similarity and the first preset coefficient.

[0038] According to a translation quality analysis device provided by the present invention, a post-translation review evaluation is obtained based on the second modification mark information and the translation quality information, including:

[0039] When the second modification marker information is the return modification marker information, the post-translation review evaluation is obtained based on the difference between the second preset coefficient and the translation quality information. The return modification marker information includes: minor modification marker information, moderate modification marker information, or severe modification marker information.

[0040] If the second modification marker information is that no modification marker information is required, the translation quality information will be used as the post-translation review and evaluation.

[0041] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the translation quality analysis method as described above.

[0042] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the translation quality analysis method as described above.

[0043] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the translation quality analysis method as described above.

[0044] This invention provides a method, apparatus, device, medium, and program product for analyzing translation quality. It analyzes translation quality information for a translation task by evaluating two aspects: modification mark quality information and post-translation quality information. Modification mark quality information is obtained through the first similarity between the computer-translated text and the standard translation. The smaller the difference between the first similarity and the modification mark information, the better the modification mark quality of the translation task; conversely, the smaller the difference, the worse the quality. Post-translation quality information is obtained by calculating the second similarity between the corresponding post-translated text and the standard translation, as well as the second modification mark information. The larger the second similarity, the higher the quality of the post-translated text; conversely, the smaller the second similarity, the lower the quality. The second modification mark information corresponds to the degree of modification in the post-translated text. The combined analysis of these two indicators yields the post-translation quality information. Finally, by analyzing the post-translation quality information and modification mark quality information, the translation quality information corresponding to the translation task can be determined, thereby helping users effectively assess the quality of the translation task and have a firm grasp on the overall project delivery quality. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0046] Figure 1 This is a schematic flowchart illustrating the translation quality analysis method described in the embodiments of this application;

[0047] Figure 2 A schematic diagram of the translation quality analysis device provided in the embodiments of this application;

[0048] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0050] Figure 1 This is a schematic flowchart of the translation quality analysis method described in the embodiments of this application, such as... Figure 1 As shown, it includes:

[0051] Step 110: Obtain translation task information, which includes at least one of the following: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and standard translation.

[0052] The translation task information described in this application embodiment refers to data obtained from historical data of completed translation tasks, including at least one of the following: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and standard translation.

[0053] The machine-translated text described in this application refers to the translation produced by specific software or machine translation.

[0054] The post-translation text described in the embodiments of this application refers to the translation that has been translated again by a translator based on the machine translation text. Therefore, it can be understood that the accuracy of the post-translation text is usually higher than that of the machine translation text.

[0055] The first modification marker information of the machine-translated text described in the embodiments of this application refers to the markers made by the translator to the machine-translated text, indicating the extent to which the translator believes the current machine-translated text needs to be modified.

[0056] The second modification marker information of the machine-translated text described in the embodiments of this application refers to the markers made by the post-translation reviewers to the post-translation text, indicating the extent to which the post-translation reviewers believe the current post-translation text needs to be modified.

[0057] Specifically, the first and second modification marker information have four possible outcomes: no modification required, minor modification, moderate modification, and severe modification. These four outcomes are represented by numerical values: 0 corresponds to no modification required, 1 to minor modification, 2 to moderate modification, and 3 to severe modification.

[0058] The standard translation described in this application refers to a translation with extremely high accuracy. It can be a translated translation that the quality control personnel deem acceptable and approves without any modifications, or a translated translation that the quality control personnel modify and that does not require further modification, or a translated translation that is translated directly by the quality control personnel without prior machine translation.

[0059] Specifically, quality control personnel refer to those responsible for ensuring the quality of translations; they are the final personnel in the translation process responsible for controlling the quality of the translated text. Understandably, quality control personnel are those widely recognized as having stronger translation skills than other translators.

[0060] Specifically, the translation task process consists of four steps: translators mark the machine-translated text with the first modification information, translators translate the machine-translated text again to generate a post-translated text, the post-translated text is reviewed by the reviewers, and the quality control personnel confirm the reviewed post-translated text again.

[0061] Step 120: Determine the modification mark quality information of the translation task information based on the first similarity between the machine-translated text and the standard translation;

[0062] The first similarity described in the embodiments of this application refers to the similarity between the machine-translated text and the standard translation.

[0063] Specifically, the first similarity is calculated from the data according to the first similarity calculation formula, and the data includes: Levenshtein Distance (edit distance), Jaccard Distance (Jaccard similarity coefficient), and Bilingual evaluation understudy (bilingual evaluation replacement distance).

[0064] Specifically, edit distance is a quantitative measure of the difference between two strings (e.g., English words) corresponding to machine-translated text and standard translation. It measures the minimum number of processing steps required to transform one string into the other. Edit distance can be used in natural language processing.

[0065] Specifically, the Jaccard similarity coefficient is used to compare the similarity and differences between machine-translated text and standard translation. The higher the Jaccard similarity coefficient value, the greater the similarity between the machine-translated text and the standard translation.

[0066] Specifically, the bilingual evaluation substitution distance is used to analyze the similarity between the machine-translated text and the reference translation. The number of words appearing simultaneously in both the machine-translated text and the reference translation is counted. Finally, the number of matching words is divided by the number of words in the machine-translated text to determine the similarity between the machine-translated text and the reference translation.

[0067] In this embodiment of the application, the above three distances are used as the basis for calculating the first similarity calculation model, thereby obtaining the first similarity between the machine-translated text and the standard translation.

[0068] Specifically, the formula for calculating the first similarity is as follows:

[0069]

[0070] Among them, ja distance Let le be the Jaccard similarity coefficient. distance To edit distance, b distance Replace the distance for bilingual evaluations; the distance value should not exceed 1.

[0071] The modification mark quality information described in this application refers to information obtained by analyzing the behavior of translators in modifying machine-translated text marks, which serves as the basis for quality analysis of the behavior. The modification mark quality information is specifically expressed as a precise numerical value.

[0072] Step 130: Determine the post-translation quality information of the translation task information based on the second similarity between the post-translated text and the standard translation and the second modification marker information;

[0073] The second similarity described in the embodiments of this application refers to the similarity between the translated text and the standard translation.

[0074] Specifically, the second similarity is calculated from the data according to the second similarity calculation model, wherein the data includes: edit distance, Jaccard similarity coefficient, and bilingual evaluation replacement distance.

[0075] In this embodiment of the application, the above three distances are used as the basis for calculating the second similarity calculation model, thereby obtaining the second similarity between the translated text and the standard translation.

[0076] Specifically, the formula for calculating the second similarity is as follows:

[0077]

[0078] Among them, ja distance Let le be the Jaccard similarity coefficient. distance To edit distance, b distance Replace the distance for bilingual evaluations; the distance value should not exceed 1.

[0079] The post-translation quality information described in this embodiment refers to information derived from quality analysis of the translator's further translation and modification of the machine-translated text. This post-translation quality information allows for a relatively intuitive evaluation of the translator's work quality. The second similarity score, used to analyze the similarity between the post-translated text and the standard translation, together with the second modification marker information, serves as a reference standard for post-translation quality.

[0080] Specifically, the reference standard refers to the change in post-translation quality information as the second similarity and second modification marker information change. If the second similarity is low, and the second modification marker information exists and has a low value, it indicates a high post-translation quality; if the second similarity is high, and the second modification marker information exists and has a high value, it indicates a poor post-translation quality. If the second similarity is low, but the second modification marker information is absent, then only the second similarity is used as the reference standard for post-translation quality; a high second similarity indicates a high post-translation quality, and vice versa.

[0081] Step 140: Determine the translation quality information corresponding to the translation task information based on the post-translation quality information and the modification mark quality information.

[0082] The translation quality information described in this embodiment refers to the analysis result obtained by comprehensively analyzing the post-translation quality information and the modification mark quality information. It is an indicator for analyzing and evaluating the overall completion quality of the translation task. The translation quality information value is the sum of 20% of the modification mark quality information value and 80% of the post-translation quality information value. The modification mark quality information corresponds to the machine-translated text, and the post-translation quality information corresponds to the post-translated text. The translation quality information value is calculated every 300 characters of the translated text. If the unit does not exceed 300 characters, the proportion of the actual number of characters in 300 characters is calculated to convert the translation quality information value in the unit. The translation quality information is obtained by accumulating and analyzing this information.

[0083] In this embodiment, a series of interconnected analytical methods are used to analyze translation quality. These methods determine modification marker quality information based on the first similarity between the machine-translated text and the standard translation, and determine the post-translation quality information based on the second similarity between the post-translated text and the standard translation and the second modification marker information. By comprehensively analyzing the post-translation quality information and the modification marker quality information, the translation quality information for the corresponding translation task can be effectively determined, thereby helping users effectively assess the quality of the translation task and have a firm grasp on the overall project delivery quality.

[0084] Optionally, based on a first similarity between the machine-translated text and the standard translation, modification marker quality information of the translation task information is determined, including:

[0085] The modification category of the machine-translated text is determined based on the first similarity between the machine-translated text and the standard translation.

[0086] Based on the modification category and the first modification marker information, the modification marker quality information of the translation task information is determined.

[0087] The modification categories of machine-translated text described in this application refer to the degree of modification required by the machine-translated text as determined by a first similarity score, and are categorized as: no modification required, minor modification, moderate modification, and severe modification. A higher similarity score indicates higher translation quality and less modification required; conversely, a lower similarity score indicates lower translation quality and more modification required. Each modification type corresponds to a modification type value.

[0088] Specifically, if the first similarity is 1, the corresponding modification category value is 0, and the corresponding modification type is no modification required; if the first similarity value is in the range of [0.8, 1), the corresponding modification category value is 1, and the corresponding modification type is mild modification; if the first similarity value is in the range of [0.5, 0.8), the corresponding modification category value is 2, and the corresponding modification type is moderate modification; if the first similarity value is no more than 0.5, the corresponding modification category value is 3, and the corresponding modification type is severe modification.

[0089] The modification mark quality information described in this embodiment is an indicator for evaluating the accuracy of the translator's modifications to the machine-translated text marks. The accuracy of the translator's modifications is measured by the degree of difference between the modification type and the first modification mark information. A greater difference between the modification type and the first modification mark information indicates that the translator's modifications to the machine-translated text marks are less accurate, and the modification mark quality is lower; conversely, a smaller difference indicates that the translator's modifications to the machine-translated text marks are more accurate, and the modification mark quality is higher.

[0090] In this embodiment, the modification tag quality information is analyzed by the degree of difference between the modification category and the first modification tag information. If the difference between the modification category and the first modification tag information is small, it indicates that the modification tag quality is high; otherwise, it indicates that the modification tag quality is poor. This modification tag quality information allows users to intuitively and clearly understand the modification tag quality.

[0091] Optionally, based on the second similarity between the translated text, the standard translation, and the second modification marker information, the post-translation quality information of the translation task information is determined, including:

[0092] Based on the second similarity between the translated text and the standard translation, the translation quality information of the translated text is determined;

[0093] Based on the second modification mark information and the translation quality information, a post-translation review evaluation is obtained;

[0094] Based on the translated text quality information and the post-translation review and evaluation, the post-translation quality information of the translation task information is determined.

[0095] The post-translation review and evaluation described in this application refers to the analysis and evaluation of the review quality by the post-translation reviewers after reviewing the translated text and providing second modification mark information, combined with the translated text quality information.

[0096] Specifically, the smaller the difference between the second modification mark information and the translated text quality information, the better the post-translation review evaluation; conversely, the worse the post-translation review evaluation.

[0097] Specifically, if the second modification marker corresponds to any of the following: minor modification, moderate modification, or severe modification, the post-translation reviewer may decide whether the original translator of the post-translation should revise it again. If the post-translation reviewer determines that the post-translation needs to be translated again by another translator, the original translator of the post-translation cannot translate the post-translation again.

[0098] Specifically, in this embodiment, the post-translation quality information of the translation task information is analyzed from two aspects: the translation quality information of the translated text and the post-translation review evaluation. If both the translation quality information and the post-translation review evaluation exist, the post-translation quality information value is 0.5 * translation quality information value + 0.5 * post-translation review evaluation value; if the translation quality information exists but the post-translation review evaluation does not exist, the post-translation quality value is the translation quality information value; if the post-translation review evaluation exists but the translation quality information does not exist, the post-translation quality value is the post-translation review evaluation value; if neither the translation quality information nor the post-translation review evaluation exists, the post-translation quality information value is 50. The range of the post-translation quality information value is [0, 100].

[0099] In this embodiment, a higher post-translation quality information value indicates higher post-translation quality, and vice versa. The post-translation quality information of the translation task information is analyzed from two aspects: the quality of the translated text and the post-translation review and evaluation. This comprehensive and multi-faceted analysis yields the post-translation quality information, effectively assessing the overall quality and allowing users to better understand and grasp the overall project completion quality.

[0100] Optionally, based on the second similarity between the translated text and the standard translation, the translation quality information of the translated text is determined, including:

[0101] The second similarity between the translated text and the standard translation is determined based on the edit distance, Jaccard distance, and bilingual evaluation substitution distance between the translated text and the standard translation.

[0102] The translation quality information of the translated text is determined based on the second similarity and the first preset coefficient.

[0103] The first preset coefficient described in this embodiment refers to a parameter defined by the user to expand the second similarity value to a user-defined range. If the user defines a percentage system, the first preset coefficient is 100; if the user defines a ten-point system, the first preset coefficient is 10.

[0104] The translation quality information described in this embodiment refers to information used to measure the quality of the translated text through analysis. The translation quality information is represented numerically, and the numerical value is the product of the first preset coefficient and the second similarity. A higher numerical value indicates a higher quality translated text, and vice versa.

[0105] In this embodiment, the translation quality information of the post-translated text is determined by the second similarity and the first preset coefficient between the post-translated text and the standard translation. This reflects the quality of the translation and allows users to intuitively understand the completed quality of the post-translated text, which is helpful for users to further improve their translation quality.

[0106] Optionally, based on the second modification marker information and the translation quality information, a post-translation review evaluation is obtained, including:

[0107] When the second modification marker information is the return modification marker information, the post-translation review evaluation is obtained based on the difference between the second preset coefficient and the translation quality information. The return modification marker information includes: minor modification marker information, moderate modification marker information, or severe modification marker information.

[0108] If the second modification marker information is that no modification marker information is required, the translation quality information will be used as the post-translation review and evaluation.

[0109] The second modification marker information described in this application embodiment refers to the case where the second modification marker information is any one of minor modification, moderate modification, or severe modification.

[0110] The situation described in the embodiments of this application where the second modified marker information is the case where the second modified marker information does not need to be modified refers to the case where the second modified marker information does not need to be modified.

[0111] The second preset coefficient described in this embodiment refers to the maximum range parameter of the post-translation evaluation value defined by the user. If the user defines 100 as the highest score, then the second preset coefficient is 100; if the user defines 10 as the highest score, then the first preset coefficient is 10.

[0112] In this embodiment, by considering two cases—one where the second modification marker information is a returned modification marker information, and the other where the second modification marker information is a marker information that does not need modification—the post-translation review evaluation can be effectively determined, allowing users to clearly understand the post-translation review evaluation status, thereby helping users improve the quality of the post-translation review work for the translation task.

[0113] The translation quality analysis apparatus provided by the present invention will be described below. The translation quality analysis apparatus described below can be referred to in correspondence with the translation quality analysis method described above.

[0114] Figure 2 This is a schematic diagram of the structure of the statement analysis device provided in the embodiments of this application, such as... Figure 2 As shown, the system includes: an acquisition module 210, a first determination module 220, a second determination module 230, and a third determination module 240; wherein, the acquisition module 210 is used to acquire translation task information, the translation task information including at least one of: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and a standard translation; wherein, the first determination module 220 is used to determine the modification mark quality information of the translation task information based on the first similarity between the machine-translated text and the standard translation; the second determination module 230 is used to determine the post-translation quality information of the translation task information based on the second similarity between the post-translated text, the standard translation, and the second modification mark information; the third determination module 240 is used to determine the translation quality information corresponding to the translation task information based on the post-translation quality information and the modification mark quality information.

[0115] Optionally, the first determining module is specifically used for:

[0116] The modification category of the machine-translated text is determined based on the first similarity between the machine-translated text and the standard translation.

[0117] Based on the modification category and the first modification marker information, the modification marker quality information of the translation task information is determined.

[0118] In this embodiment, the quality information of the modification mark is analyzed by the difference between the modification category and the first modification mark information. If the difference between the modification category and the first modification mark information is small, it indicates that the quality of the modification mark is high, and vice versa. This can effectively ensure the accuracy of the calculated quality information of the modification mark, so that users can intuitively and clearly understand the quality information of the modification mark.

[0119] Optionally, the second determining module is specifically used for:

[0120] Based on the second similarity between the translated text and the standard translation, the translation quality information of the translated text is determined;

[0121] Based on the second modification mark information and the translation quality information, a post-translation review evaluation is obtained;

[0122] Based on the translated text quality information and the post-translation review and evaluation, the post-translation quality information of the translation task information is determined.

[0123] In this embodiment, a higher post-translation quality information value indicates higher post-translation quality, and vice versa. The post-translation quality information of the translation task information is analyzed from two aspects: the quality of the translated text and the post-translation review and evaluation. This comprehensive and multi-faceted analysis yields the post-translation quality information, effectively assessing the overall quality and allowing users to better understand and grasp the overall project completion quality.

[0124] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 3 As shown, the electronic device may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other via the communication bus 340. The processor 310 can invoke logical instructions in the memory 330 to execute a translation quality analysis method, which includes: acquiring translation task information, the translation task information including at least one of: machine-translated text, first modification marker information of the machine-translated text, post-translated text, second modification marker information of the post-translated text, and a standard translation.

[0125] Based on the first similarity between the machine-translated text and the standard translation, the modification mark quality information of the translation task information is determined;

[0126] Based on the second similarity between the post-translated text and the standard translation, and the second modification marker information, the post-translation quality information of the translation task information is determined;

[0127] The translation quality information corresponding to the translation task information is determined based on the post-translation quality information and the modification mark quality information.

[0128] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0129] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, wherein when the computer program is executed by a processor, the computer is able to execute the translation quality analysis method provided by the above methods, the method comprising: acquiring translation task information, the translation task information comprising at least one of: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and standard translation;

[0130] Based on the first similarity between the machine-translated text and the standard translation, the modification mark quality information of the translation task information is determined;

[0131] Based on the second similarity between the post-translated text and the standard translation, and the second modification marker information, the post-translation quality information of the translation task information is determined;

[0132] The translation quality information corresponding to the translation task information is determined based on the post-translation quality information and the modification mark quality information.

[0133] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a translation quality analysis method provided by the above methods, the method comprising: acquiring translation task information, the translation task information including at least one of: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and standard translation;

[0134] Based on the first similarity between the machine-translated text and the standard translation, the modification mark quality information of the translation task information is determined;

[0135] Based on the second similarity between the post-translated text and the standard translation, and the second modification marker information, the post-translation quality information of the translation task information is determined;

[0136] The translation quality information corresponding to the translation task information is determined based on the post-translation quality information and the modification mark quality information.

[0137] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. 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 the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0138] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for analyzing translation quality, characterized in that, include: Obtain translation task information, which includes at least one of the following: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and standard translation; the post-translated text is a translation that has been further translated by translators based on the machine-translated text. Based on the first similarity between the machine-translated text and the standard translation, the modification mark quality information of the translation task information is determined; Based on the second similarity between the post-translated text and the standard translation, and the second modification marker information, the post-translation quality information of the translation task information is determined; The translation quality information corresponding to the translation task information is determined based on the post-translation quality information and the modification mark quality information. The first modification marker information and the second modification marker information respectively include: no modification, slight modification, moderate modification, or severe modification; wherein, if the first similarity is 1, the corresponding modification type is no modification; if the first similarity value is in the range of [0.8, 1), the corresponding modification type is slight modification; if the first similarity value is in the range of [0.5, 0.8), the corresponding modification type is moderate modification; if the first similarity value does not exceed 0.5, the corresponding modification type is severe modification.

2. The translation quality analysis method according to claim 1, characterized in that, Based on the first similarity between the machine-translated text and the standard translation, the modification marker quality information of the translation task information is determined, including: The modification category of the machine-translated text is determined based on the first similarity between the machine-translated text and the standard translation. Based on the modification category and the first modification marker information, the modification marker quality information of the translation task information is determined.

3. The method for analyzing translation quality according to claim 1, characterized in that, Based on the second similarity between the post-translated text and the standard translation, and the second modification marker information, the post-translation quality information of the translation task information is determined, including: Based on the second similarity between the translated text and the standard translation, the translation quality information of the translated text is determined; Based on the second modification mark information and the translation quality information, a post-translation review evaluation is obtained; Based on the translated text quality information and the post-translation review and evaluation, the post-translation quality information of the translation task information is determined.

4. The translation quality analysis method according to claim 3, characterized in that, Based on the second similarity between the translated text and the standard translation, the translation quality information of the translated text is determined, including: The second similarity between the translated text and the standard translation is determined based on the edit distance, Jaccard distance, and bilingual evaluation substitution distance between the translated text and the standard translation. The translation quality information of the translated text is determined based on the second similarity and the first preset coefficient.

5. The translation quality analysis method according to claim 3, characterized in that, Based on the second modification marker information and the translation quality information, a post-translation review and evaluation is obtained, including: When the second modification marker information is the return modification marker information, the post-translation review evaluation is obtained based on the difference between the second preset coefficient and the translation quality information. The return modification marker information includes: minor modification marker information, moderate modification marker information, or severe modification marker information. If the second modification marker information is that no modification marker information is required, the translation quality information will be used as the post-translation review and evaluation.

6. A translation quality analysis device, characterized in that, include: The acquisition module is used to acquire translation task information, which includes at least one of the following: machine-translated text, first modification mark information of the machine-translated text, post-translated text, second modification mark information of the post-translated text, and standard translation; the post-translated text is a translation that has been translated again by translators based on the machine-translated text. The first determining module is used to determine the modification mark quality information of the translation task information based on the first similarity between the machine-translated text and the standard translation; The second determining module is used to determine the post-translation quality information of the translation task information based on the second similarity between the post-translated text and the standard translation and the second modification mark information; The third determining module is used to determine the translation quality information corresponding to the translation task information based on the post-translation quality information and the modification mark quality information; The first modification marker information and the second modification marker information respectively include: no modification, slight modification, moderate modification, or severe modification; wherein, if the first similarity is 1, the corresponding modification type is no modification; if the first similarity value is in the range of [0.8, 1), the corresponding modification type is slight modification; if the first similarity value is in the range of [0.5, 0.8), the corresponding modification type is moderate modification; if the first similarity value does not exceed 0.5, the corresponding modification type is severe modification.

7. The translation quality analysis device according to claim 6, characterized in that, The first determining module is specifically used for: The modification category of the machine-translated text is determined based on the first similarity between the machine-translated text and the standard translation. Based on the modification category and the first modification marker information, the modification marker quality information of the translation task information is determined.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the translation quality analysis method as described in any one of claims 1 to 5.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the translation quality analysis method as described in any one of claims 1 to 5.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the translation quality analysis method as described in any one of claims 1 to 5.

Images