CN111680524B - Human-machine feedback translation method and system based on inverse matrix analysis - Google Patents
Human-machine feedback translation method and system based on inverse matrix analysis Download PDFInfo
- Publication number
- CN111680524B CN111680524B CN202010505910.0A CN202010505910A CN111680524B CN 111680524 B CN111680524 B CN 111680524B CN 202010505910 A CN202010505910 A CN 202010505910A CN 111680524 B CN111680524 B CN 111680524B
- Authority
- CN
- China
- Prior art keywords
- translation
- machine
- semantic
- matrix
- engine
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/40—Processing or translation of natural language
- G06F40/58—Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/40—Processing or translation of natural language
- G06F40/42—Data-driven translation
- G06F40/44—Statistical methods, e.g. probability models
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Machine Translation (AREA)
Abstract
The invention provides a human-machine feedback translation method based on inverse matrix analysis, a human-machine co-translation system with feedback adjustment and a computer readable storage medium for realizing the method. The method comprises the steps of receiving a document to be translated, carrying out semantic recognition on the document to be translated, and utilizing a combined machine translation tool to translate the document to be translated to generate a similarity matrix and a difference matrix so as to select a reverse translation engine to compare differences for man-machine co-translation. By adopting the technical scheme of the invention, the accurate opportunity of manual translation intervention can be provided, so that the translation efficiency can be ensured for large-scale corpus translation and translation occasions with higher accuracy requirements, and the accuracy of translation can be ensured, thereby realizing the requirement of large-scale semantic translation while ensuring the accuracy.
Description
Technical Field
The invention belongs to the technical field of translation, and particularly relates to a human-machine feedback translation method based on inverse matrix analysis, a human-machine co-translation system with feedback adjustment and a computer readable storage medium for realizing the method.
Background
Machine translation is a process of translating one natural language into another natural language by using a computer, and the basic flow is roughly divided into three blocks: preprocessing, core translation and post-processing. The pretreatment is to normalize language characters, divide overlong sentences into a plurality of short sentences through punctuation marks, filter some language words and characters irrelevant to meaning, and normalize some numbers and places with nonstandard expression into sentences conforming to the specifications. The core translation module is a process of translating input character units and sequences into target language sequences, and is the most key place in machine translation. The post-processing module is used for splicing the conversion of the upper and lower cases of the translation result and the modeling unit and processing the special symbols, so that the translation result accords with the reading habit of people.
The dream of achieving high quality machine translation has existed for many years and many scientists have contributed to this dream in their own time and heart. From early rule-based machine translation to nerve machine translation widely used today, the level of machine translation is continuously increasing, and the basic application requirements of many scenes can be met.
In the age of large-scale translation engineering, the use of machine translation tools is unavoidable. However, the limitations of machine translation make it impossible to completely replace manual translation. Nevertheless, if the correct translation tool is selected, the translation efficiency will be greatly improved, which is undeniable.
The prior art has presented translation tools and machine translation engines between various languages, including various optimized machine translation schemes developed with artificial intelligence, big data, deep learning techniques.
Two of the most important machine translation modes are currently: rule methods and statistical methods. The rule method (rule based machine translation, RBMT) analyzes the text according to language rules and translates it by means of a computer program. Most commercial machine translation systems employ rule methods. The statistical method (statistical machine translation, SMT) constructs a statistical translation model (vocabulary, comparison or language mode) by performing statistical analysis on a large number of parallel corpora, and then uses the model to translate, generally selects the entry with the highest occurrence probability in statistics as translation, and the probability algorithm is based on bayesian theorem. It is assumed that an english sentence a is to be translated into chinese, and that all chinese sentences B are potential translations of a that are possible or not possible. Pr (A) is the probability of the appearance of a similar A expression, pr (B|A) is the probability of the appearance of A translation into B. Finding the maximum value of the two parameters can reduce the range of sentences and corresponding translation search, thereby finding out the most suitable translation. SMT is classified into two types according to the difference of the text analysis degree level: word-based SMT and phrase-based SMT, the latter being currently in common use, google being the only thing that is used. The translation text is automatically divided into word sequences with fixed length, and each word sequence is subjected to statistical analysis in a corpus so as to find the translation with the highest occurrence probability.
The Chinese patent application with the application number of CN201910772953.2 provides a method and a system for evaluating a machine translation engine based on sentence pairs, wherein the method comprises the steps of scoring a plurality of dimensions of each machine translation engine according to the selected language pairs and the sentence field, weighting and summing the scores to obtain a weighted sum value of each machine translation engine on the sentence, and selecting a machine translation engine with the highest weighted sum value to output the translation result of the sentence, so that the whole translation text is obtained through integration. By the method, the automatic optimal service of the machine translation engine can be provided for the user in various complicated machine translation engines with uneven translation quality, good field and different language pairs, so that the user can obtain the best machine translation engine service at present when translating texts such as documents, the translation efficiency is improved, the follow-up workload of the user is reduced, and the high-quality machine translation service is provided.
The Chinese patent application with the application number of CN201810063565.2 provides a Machine synchronous translation device and a Machine synchronous translation method in the professional field based on deep learning, which promote simultaneous interpretation to develop from a translator simultaneous transmission (Human SI) to a Machine assisted simultaneous transmission (Computer-aid SI) and then to a final Machine interpretation (Machine SI). On the one hand, the method solves the difficulty that a translator cannot understand, remember and interpret in the simultaneous interpretation site, improves the bilingual conversion accuracy of terms and inherent expressions, on the other hand, the translator can edit after translating on line in real time according to machine translation, improves the information quantity of translated languages, and can replace the translator to realize the quasi-real-time simultaneous interpretation function within one second in some scenes.
The Chinese patent application with the application number of CN201710203439.8 provides a multi-language intelligent preprocessing real-time statistical machine translation system which can translate sentences and chapters of one language into another language in real time, the system can translate the sentences completely, express correctly, translate text languages with punctuation marks, and translate voices with noise in the sentences without segmentation and segmentation, wherein the sentences may be incomplete without punctuation marks; the translation accuracy of the words and phrases with small probability is improved, namely, the words with small probability such as numbers, dates, time, URLs and the like are respectively marked and translated with priority; the pretreatment module can carry out standardization treatment on the input sentences; the post-processing module can improve the fluency of the translation result.
While the impact of machine translation on the translation industry is fatal, the average translator who undertakes the simple translation task of not requiring accuracy will be thoroughly replaced by a machine. However, for specialized translation work, it appears that machines are not adequate. But the reason for the inability to translate is not how hard it is in its content. In fact, the difficulty of specialized domain translation is not high, and in addition to the proprietary vocabulary commonly used in the industry, syntactically strict and standardized writing is often used in each domain, but specialized translation requires extremely high accuracy, and a machine cannot bear serious consequences caused by errors. Many times, some of the very specialized documents are often translated in person by professionals in their field, which also explains why some of the concepts are extremely robust, logical and unaesthetic.
However, in the age of large-scale translation engineering, the efficiency requirement is obviously not satisfied by only manual translation, and the accuracy requirement is not satisfied by only machine translation. Regardless of the improvements in the scheme of machine translation or computer-aided translation, manual editing proofreading is indispensable even for translation. Furthermore, a translator typically selects a translation tool based on experience, usage habits, and characteristics of the translation tool, and once selected, is essentially unchanged.
However, it has been found in practice that the cores of different translation tools are different, often giving different results for the same text to be translated; in addition, the accuracy of the results of the same translation tool is different for the same piece of data to be translated, wherein different sub-parts. Existing translators are often quite confused about this and wander back and forth among a variety of translation tools, tiring in selecting translation tools and collating translation results. Such a variety, the translation tools originally used to increase translation efficiency, in turn, are cumbersome to drag down the translation tempo.
It can be seen how to balance the working time of manual editing translation and machine translation, when manual translation intervenes, in what way and can meet the needs of large-scale semantic translation while guaranteeing accuracy, the prior art does not give an effective solution.
Disclosure of Invention
In order to solve the technical problems, the invention provides a human-machine feedback translation method based on inverse matrix analysis, a human-machine co-translation system with feedback adjustment and a computer readable storage medium for realizing the method. The method comprises the steps of receiving a document to be translated, carrying out semantic recognition on the Tobedoc of the document to be translated, and utilizing a combined machine translation tool to translate the document to be translated to generate a similarity matrix and a difference matrix so as to select a reverse translation engine to compare differences for man-machine co-translation. By adopting the technical scheme of the invention, the accurate opportunity of manual translation intervention can be provided, so that the translation efficiency can be ensured for large-scale corpus translation and translation occasions with higher accuracy requirements, and the accuracy of translation can be ensured, thereby realizing the requirement of large-scale semantic translation while ensuring the accuracy.
Specifically, in a first aspect of the present invention, a human-machine feedback translation method based on inverse matrix analysis is provided, the method comprising the steps of:
s1: receiving a to-be-translated document Tobedoc;
s2: carrying out semantic recognition on the Tobedoc of the document to be translated to obtain a semantic sentence subset TobeSen taking a semantic sentence as a unit, wherein the semantic sentence subset TobeSen consists of a plurality of semantic sentences Seni=
1,2, … … are positive integers;
s3: for each semantic sentence Seni in the semantic sentence subset TobeSen, performing the following translation processing procedure until all semantic sentences are translated:
f001: for each semantic sentence Seni, translating the semantic sentence by using a combined machine translation tool, and outputting at least three target translations Y1, Y2 and Y3;
the combined machine translation tool comprises at least three machine translation engines, and the three target translations Y1, Y2 and Y3 are respectively output by the three machine translation engines;
f002: respectively calculating the similarity Sij and the difference Dij between every two of the three target translations Y1, Y2 and Y3, so as to obtain a similarity matrix Sm and a difference matrix Dm;
f003: based on the similarity matrix and the difference matrix, selecting one of three target translations Y1, Y2 and Y3 as a translation result of each semantic sentence Seni;
f004: selecting a machine translation engine corresponding to one of the three target translations Y1, Y2 and Y3 as a reverse translation engine based on the similarity matrix and the difference matrix, and reversely translating the three target translations Y1, Y2 and Y3 by the reverse translation engine;
f005: and comparing the difference degree of the reverse translation result and the semantic sentence Seni and displaying the result on a man-machine co-translation interface.
More specifically, as one of the key technical means of the present invention, the step F003 specifically includes:
searching a maximum element value Smax in the similarity matrix Sm;
searching a minimum element value Dmin in the difference matrix Dm;
and taking the common target translation corresponding to the maximum element value Smax and the minimum element value Dmin as a translation result of the semantic sentence Seni.
More specifically, as one of the key technical means of the present invention, the step F004 specifically includes:
searching a minimum element value Smin in the similarity matrix Sm;
searching a maximum element value Dmax in the difference matrix Dm;
and taking a machine translation engine which outputs the common target translation corresponding to the minimum element value Smin and the maximum element value Dmax as the reverse translation engine.
On the basis, the technical scheme of the invention further comprises the following steps: the weight of the similarity matrix element corresponding to the reverse translation engine is reduced,
and the weight of the difference matrix element corresponding to the machine translation engine outputting the common target translation corresponding to the maximum element value Smax and the minimum element value Dmin is improved.
As a further preference, the three machine translation engines of the combined machine translation tool comprise ICAT, TRADOS, LINGOES.
It should be noted that, the reverse translation refers to translating the target translation into the language of the document to be translated TobeDoc through one of the three machine translation engines.
In a second aspect of the present invention, there is also provided a man-machine co-translation system with feedback adjustment, the man-machine co-translation system comprising a combined machine translation tool and man-machine co-translation interface, characterized in that:
the man-machine co-translation system also comprises a semantic recognition engine and a matrix generation engine, and is used for realizing the man-machine feedback translation method.
The above method of the present invention may be implemented by a program code in the form of computer instructions, and thus the present invention also provides a computer-readable storage medium having stored thereon computer-executable instructions for implementing the aforementioned human-machine feedback translation method based on inverse matrix analysis by a communication terminal comprising a memory and a processor.
Further advantages of the invention will be further elaborated in the description section of the embodiments in connection with the drawings.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a main flow chart of a human-machine feedback translation method according to an embodiment of the present invention.
Fig. 2-4 are further embodiments of the method described in fig. 1.
Fig. 5 is a diagram of a computer system architecture for implementing the methods described in fig. 1-4.
Detailed Description
The invention will be further described with reference to the drawings and detailed description.
Referring to fig. 1, a main flow chart of a man-machine feedback translation method of an embodiment of the present invention is shown.
In fig. 1, the method mainly comprises three main stages S1-S3:
s1: receiving a to-be-translated document Tobedoc;
s2: carrying out semantic recognition on the Tobedoc of the document to be translated to obtain a semantic sentence subset TobeSen taking semantic sentences as units, wherein the semantic sentence subset TobeSen consists of a plurality of semantic sentences Seni=1, 2 and … … is a positive integer;
s3: and executing a translation processing process on each semantic sentence Seni in the semantic sentence subset TobeSen until all semantic sentences are translated.
Specific implementation of the translation process is described with reference to steps F001-F005 of fig. 2:
f001: for each semantic sentence Seni, translating the semantic sentence by using a combined machine translation tool, and outputting at least three target translations Y1, Y2 and Y3;
the combined machine translation tool comprises at least three machine translation engines, and the three target translations Y1, Y2 and Y3 are respectively output by the three machine translation engines;
f002: respectively calculating the similarity Sij and the difference Dij between every two of the three target translations Y1, Y2 and Y3, so as to obtain a similarity matrix Sm and a difference matrix Dm;
f003: based on the similarity matrix and the difference matrix, selecting one of three target translations Y1, Y2 and Y3 as a translation result of each semantic sentence Seni;
f004: selecting a machine translation engine corresponding to one of the three target translations Y1, Y2 and Y3 as a reverse translation engine based on the similarity matrix and the difference matrix, and reversely translating the three target translations Y1, Y2 and Y3 by the reverse translation engine;
f005: and comparing the difference degree of the reverse translation result and the semantic sentence Seni and displaying the result on a man-machine co-translation interface.
With further reference to fig. 3-4, the step F003 specifically includes:
searching a maximum element value Smax in the similarity matrix Sm;
searching a minimum element value Dmin in the difference matrix Dm;
and taking the common target translation corresponding to the maximum element value Smax and the minimum element value Dmin as a translation result of the semantic sentence Seni.
The step F004 specifically comprises the following steps:
searching a minimum element value Smin in the similarity matrix Sm;
searching a maximum element value Dmax in the difference matrix Dm;
and taking a machine translation engine which outputs the common target translation corresponding to the minimum element value Smin and the maximum element value Dmax as the reverse translation engine.
As a non-limiting example, the similarity matrix and the difference matrix described in fig. 3-4 may be expressed as follows:
wherein s is ij Represents the similarity between Yi and Yj, d ij The degree of difference between Yi and Yj is expressed.
How to calculate the degree of similarity or the degree of similarity, there are various common methods and numerical standards in the art, the invention is not limited to this, and the finally calculated degree of similarity or degree of similarity values are normalized to be between [0,1], for example, the degree of similarity is 1 and indicates that the two are identical, and the degree of similarity is 1 and indicates that the two are completely different.
As a non-limiting example, assume smax=ds 21 ，Dmin＝d 32 If the table is 2, the common target text corresponding to Smax and Dmin is Y2, so that Y2 is used as the translation result of the semantic sentence Seni.
As a non-limiting example, assume smin=s 31 ，Dmax＝d 32 And if the common target text corresponding to Smin and Dmax is Y3, and if Y3 corresponds to a third machine-turning engine, the third machine-turning engine is used as the reverse translation engine.
In the invention, each element of the similarity matrix and the difference matrix can be provided with an adjustable weight value as a feedback loop signal, and the weight of the similarity matrix element corresponding to the reverse comparison translation engine is reduced after the target translation and the reverse comparison translation engine are selected each time; correspondingly, the weight of the difference matrix element corresponding to the machine translation engine outputting the common target translation corresponding to the maximum element value Smax and the minimum element value Dmin is promoted.
Fig. 5 is a diagram of a computer system architecture for implementing the methods described in fig. 1-4.
The computer system is a man-machine co-translation system with feedback adjustment for implementing the method described in fig. 1-4, and the man-machine co-translation system comprises a combined machine translation tool and man-machine co-translation interface, and also comprises a semantic recognition engine and a matrix generation engine.
Experimental data shows that the technical scheme of the invention can maximally utilize the machine translation result, and meanwhile, the possible erroneous translation result is manually checked to the greatest extent, so that the accuracy is improved while the efficiency is ensured.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (7)
1. A human-machine feedback translation method based on inverse matrix analysis, the method comprising the steps of:
s1: receiving a to-be-translated document Tobedoc;
s2: carrying out semantic recognition on the Tobedoc of the document to be translated to obtain a semantic sentence subset TobeSen taking semantic sentences as units, wherein the semantic sentence subset TobeSen consists of a plurality of semantic sentences Seni=1, 2 and … … is a positive integer;
s3: for each semantic sentence Seni in the semantic sentence subset TobeSen, performing the following translation processing procedure until all semantic sentences are translated:
f001: for each semantic sentence Seni, translating the semantic sentence by using a combined machine translation tool, and outputting at least three target translations Y1, Y2 and Y3;
the combined machine translation tool comprises at least three machine translation engines, and the three target translations Y1, Y2 and Y3 are respectively output by the three machine translation engines;
f002: respectively calculating the similarity Sij and the difference Dij between every two of the three target translations Y1, Y2 and Y3, so as to obtain a similarity matrix Sm and a difference matrix Dm;
f003: based on the similarity matrix and the difference matrix, selecting one of three target translations Y1, Y2 and Y3 as a translation result of each semantic sentence Seni;
f004: selecting a machine translation engine corresponding to one of the three target translations Y1, Y2 and Y3 as a reverse translation engine based on the similarity matrix and the difference matrix, and reversely translating the three target translations Y1, Y2 and Y3 by the reverse translation engine;
f005: comparing and displaying the difference degree of the reverse translation result and the semantic sentence Seni on a man-machine co-translation interface;
the step F003 specifically includes:
searching a maximum element value Smax in the similarity matrix Sm;
searching a minimum element value Dmin in the difference matrix Dm;
taking a common target translation corresponding to the maximum element value Smax and the minimum element value Dmin as a translation result of the semantic sentence Seni;
the step F004 specifically comprises the following steps:
searching a minimum element value Smin in the similarity matrix Sm;
searching a maximum element value Dmax in the difference matrix Dm;
and taking a machine translation engine which outputs the common target translation corresponding to the minimum element value Smin and the maximum element value Dmax as the reverse translation engine.
2. The human-machine feedback translation method of claim 1, wherein:
and reducing the weight of the similarity matrix element corresponding to the reverse translation engine.
3. The human-machine feedback translation method of claim 1, wherein:
and lifting the weight of the difference matrix element corresponding to the machine translation engine outputting the common target translation corresponding to the maximum element value Smax and the minimum element value Dmin.
4. The human-machine feedback translation method of claim 1, wherein:
the three machine translation engines of the combined machine translation tool include ICAT, TRADOS, LINGOES.
5. The human-machine feedback translation method of claim 1, wherein:
the reverse translation refers to translating the target translation into the language of the to-be-translated document Tobedoc through one of the three machine translation engines.
6. A man-machine co-translation system with feedback adjustment, the man-machine co-translation system comprising a combined machine translation tool and a man-machine co-translation interface, characterized in that:
the man-machine co-interpretation system further comprises a semantic recognition engine and a matrix generation engine for implementing the method of any of claims 1-5.
7. A computer readable storage medium having stored thereon computer executable program instructions, which are executed by a processor for implementing the method of any of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010505910.0A CN111680524B (en) | 2020-06-09 | 2020-06-09 | Human-machine feedback translation method and system based on inverse matrix analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010505910.0A CN111680524B (en) | 2020-06-09 | 2020-06-09 | Human-machine feedback translation method and system based on inverse matrix analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111680524A CN111680524A (en) | 2020-09-18 |
| CN111680524B true CN111680524B (en) | 2023-06-06 |
Family
ID=72453934
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010505910.0A Active CN111680524B (en) | 2020-06-09 | 2020-06-09 | Human-machine feedback translation method and system based on inverse matrix analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111680524B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112818707B (en) * | 2021-01-19 | 2024-02-27 | 传神语联网网络科技股份有限公司 | Reverse text consensus-based multi-turn engine collaborative speech translation system and method |
| CN112818704B (en) * | 2021-01-19 | 2024-04-02 | 传神语联网网络科技股份有限公司 | Multilingual translation system and method based on inter-thread consensus feedback |
| CN112818706B (en) * | 2021-01-19 | 2024-02-27 | 传神语联网网络科技股份有限公司 | Voice translation real-time dispute recording system and method based on reverse result stability |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7680646B2 (en) * | 2004-12-21 | 2010-03-16 | Xerox Corporation | Retrieval method for translation memories containing highly structured documents |
| CN106484681B (en) * | 2015-08-25 | 2019-07-09 | 阿里巴巴集团控股有限公司 | A kind of method, apparatus and electronic equipment generating candidate translation |
| KR102458244B1 (en) * | 2017-11-23 | 2022-10-24 | 삼성전자주식회사 | Machine translation method and apparatus |
| CN108804428A (en) * | 2018-06-12 | 2018-11-13 | 苏州大学 | Correcting method, system and the relevant apparatus of term mistranslation in a kind of translation |
| CN111144134B (en) * | 2019-11-27 | 2023-05-16 | 语联网（武汉）信息技术有限公司 | OpenKiwi-based automatic evaluation system for translation engine |
| CN110991193B (en) * | 2019-11-27 | 2023-06-09 | 语联网（武汉）信息技术有限公司 | OpenKiwi-based translation matrix model selection system |
| CN111027331B (en) * | 2019-12-05 | 2022-04-05 | 百度在线网络技术（北京）有限公司 | Method and apparatus for evaluating translation quality |
-
2020
- 2020-06-09 CN CN202010505910.0A patent/CN111680524B/en active Active
Non-Patent Citations (2)
| Title |
|---|
| Iterative Back-Translation for Neural Machine Translation;Vu Cong Duy Hoang;Proceedings of the 2nd Workshop on Neural Machine Translation and Generation;18-24 * |
| 基于词性及词性依存的句子结构相似度计算;蓝雁玲;陈建超;;计算机工程(第10期);53-55 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111680524A (en) | 2020-09-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111680524B (en) | Human-machine feedback translation method and system based on inverse matrix analysis | |
| CN110532573B (en) | Translation method and system | |
| CN109840331B (en) | Neural machine translation method based on user dictionary | |
| TW201918913A (en) | Machine processing and text correction method and device, computing equipment and storage media | |
| CN111680526B (en) | Man-machine interactive translation system and method based on comparison of reverse translation results | |
| Kesidis et al. | A word spotting framework for historical machine-printed documents | |
| CN111680527B (en) | Man-machine co-interpretation system and method based on dedicated machine turning engine training | |
| CN118170894B (en) | Knowledge graph question-answering method, knowledge graph question-answering device and storage medium | |
| CN112329482A (en) | Machine translation method, device, electronic equipment and readable storage medium | |
| CN112800184B (en) | Short text comment emotion analysis method based on Target-Aspect-Opinion joint extraction | |
| CN112257460A (en) | Pivot-based Hanyue combined training neural machine translation method | |
| Liu | Research on the development of computer intelligent proofreading system based on the perspective of English translation application | |
| CN109325237B (en) | Complete sentence recognition method and system for machine translation | |
| CN115048940A (en) | Chinese financial text data enhancement method based on entity word attribute characteristics and translation | |
| Zheng et al. | A study of computer aided translation based on artificial intelligence technology | |
| CN114722842A (en) | Computer artificial intelligent foreign language translation method and translation system thereof | |
| CN114528840A (en) | Chinese entity identification method, terminal and storage medium fusing context information | |
| CN111680523B (en) | Man-machine collaborative translation system and method based on context semantic comparison | |
| CN112836528A (en) | Machine translation post-editing method and system | |
| CN102135957A (en) | Clause translating method and device | |
| CN114970524B (en) | Controllable text generation method and device | |
| CN111680525B (en) | Man-machine co-translation method and system based on reverse difference recognition | |
| Granell et al. | Study of the influence of lexicon and language restrictions on computer assisted transcription of historical manuscripts | |
| CN114861628A (en) | System, method, electronic device and storage medium for training machine translation model | |
| CN115310433A (en) | Data enhancement method for Chinese text proofreading |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |