CN109408832B - Translation quality early warning method and system based on repeated sentence detection - Google Patents
Translation quality early warning method and system based on repeated sentence detection Download PDFInfo
- Publication number
- CN109408832B CN109408832B CN201811194428.9A CN201811194428A CN109408832B CN 109408832 B CN109408832 B CN 109408832B CN 201811194428 A CN201811194428 A CN 201811194428A CN 109408832 B CN109408832 B CN 109408832B
- Authority
- CN
- China
- Prior art keywords
- translation
- sentence
- current
- translated
- correction
- 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
- G06F40/51—Translation evaluation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/194—Calculation of difference between files
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
Abstract
The invention provides a translation quality early warning method and system based on repeated sentence detection. Based on the detected repeated sentences, early warning prompt is carried out on the current translation personnel, so that the translation consistency of the repeated sentences can be ensured in the translation stage. With the invention, the translator only needs to pay attention to the current translation work, and does not need to worry about whether the translation result of the translator is consistent with other people. The technical scheme of the invention can automatically detect repeated sentences and judge whether translation quality problems exist or not so as to provide early warning prompt; after the early warning prompt is sent, a corresponding correction result can be given, and the uniformity of the translation results of all repeated sentences is maintained; this process may be accomplished by computer automated replacement.
Description
Technical Field
The application relates to the technical field of translation, in particular to a translation quality early warning method and a translation quality early warning system based on repeated sentence detection.
Background
Because of the existence of the literature, certain flexibility exists in translation work in many times, and even for the same sentence, when translation is performed in different contexts, the same translator can give different translation results; it is also very common for different translators to give different translation results for the same sentence. In most cases, such as for cultural works, news manuscripts, etc., this difference does not have a negative effect, so the translator will usually allow the difference between different translation results (as long as it expresses the same cultural connotation), and does not further process this, i.e. embody the diversity of literature translations.
However, in some specific situations, the translator must ensure a high degree of accuracy and uniformity of the translation results, without allowing random variation. In these cases, the translation results of the same sentence in the same context must be unified, whether the same person translates or different persons translate, and in the final overall translation result, the translation results for the same sentence should also be kept consistent to a considerable extent to show the seriousness and fairness of the translation results. These include mainly documents related to law, foreign contracts, government notices/news manuscripts, professional textbooks, etc. The foreign contract contains a large number of repeated template sentences, and legal documents such as foreign patents have a large number of repeated sentences due to the writing characteristics, and if the translation consistency of the same repeated sentences before and after cannot be maintained, the authority of the foreign contracts is greatly damaged.
Unfortunately, in the current translation work, it is very difficult for the translator to ensure that the translation of the same sentence is consistent from the beginning to the end, aiming at the translation requirement of the specific occasion, at least the following aspects are included:
(1) When the translation is not completed completely, it cannot be predicted whether the same sentence exists after the current sentence;
(2) When the translation amount is huge, the one-to-one comparison is not imaginable for the translator whether the same sentence exists in the past;
(3) When the translation amount is huge, a team is generally formed by a plurality of different translators to jointly complete the translation, each translator independently translates, and whether other people translate the same sentence or not cannot be known at all during the translation.
One potential idea for this problem is to search for possible identical sentences after the translation work is completed, then determine whether the translations of the identical sentences are consistent, and if not, manually correct the sentences one by one. It is conceivable that this process will be time consuming and laborious, and that as the amount of translation increases, the efficiency of translation will decrease drastically, even resulting in the correction process not being completed.
Therefore, there is a need to propose a solution to the above-mentioned problems in the current translation process.
Disclosure of Invention
The technical solution of the present invention solves the above-mentioned problems at least from the following aspects.
In a first aspect of the present invention, a translation quality early warning method based on repeated sentence detection is provided, and early warning prompt is performed on current translators based on detected repeated sentences, so that translation consistency of the repeated sentences can be ensured in a translation stage.
The method comprises the following steps:
(1) Uploading the current sentence to be translated and the current translated sentence.
Preferably, in the method, the step of uploading further includes a step of determining whether the current sentence to be translated and/or the current translation sentence need to be uploaded.
In order to save memory space and subsequent detection efficiency, not all sentences in the translation process need to be uploaded. The inventors have found that there is less likelihood of duplication for most sentences, and that the translation results for most sentences are relatively unique and that different translation results are less likely to occur. Therefore, a judging step is set for judging whether to upload the current sentence, and the judging basis can be determined by the current translator or set by a system setting parameter interface.
(2) Repeating the steps of sentence waiting and/or sentence translation detection.
Aiming at a current sentence to be translated uploaded by a current translator, detecting whether repeated sentences to be translated exist in the corpus;
and/or;
detecting whether repeated translation sentences exist in the corpus aiming at the current translation sentences uploaded by the current translation personnel;
(3) And a translation quality early warning step.
And if repeated sentences to be translated and/or repeated translated sentences are detected, early warning is carried out on the current translation quality.
(4) And correcting the translation result.
And if the translation quality early warning prompt is received, correcting the current translation result according to the actual situation. The purpose of the correction is to ensure that the translation results of all repeated sentences are consistent.
Since the early warning prompt shows that different translation results exist, the different translation results must be corrected to a consistent unified result according to a certain standard.
As another innovative point of the present invention, the above correction is performed in the following manner.
Automatic correction: and automatically completing correction by adopting an artificial intelligence system.
The artificial intelligence system automatically selects an optimal translation result as a consistent unified result according to machine learning, and comprises the following steps: automatically counting translation results of the same sentence in the historical translation corpus, and selecting the translation result with the largest selected number as a uniform result;
cross-alignment correction: displaying all the translation results aiming at the repeated sentences at present, allowing all the current translators to cross-compare and vote, and selecting the translation result with the largest voting number as a uniform result;
expert correction: the expert provides the translation result of the current repeated sentence as a consistent unified result.
In expert correction, the real-time online correction of the expert can be realized, and the current sentences to be corrected can be submitted, and unified correction can be performed when a certain number of sentences to be corrected are reached or the expert is online.
Through the mode, the translation quality can be timely found and early-warned, and quality correction is completed in real time.
Meanwhile, according to the above process, it can be seen that the translation quality early warning method of the present invention does not affect the normal translation operation, and does not require the translator to stop the current translation operation (even if the step 4 selects "cross comparison correction", only one selection voting operation is completed on the voting interface of the side, and the current translation operation process is not interrupted).
By adopting the method, the translator only needs to pay attention to the current translation work, and does not need to worry about whether the translation result of the translator is consistent with other people. As described above, the technical scheme of the invention can automatically detect repeated sentences and judge whether translation quality problems exist so as to provide early warning prompt; after the early warning prompt is sent, a corresponding correction result can be given, and the uniformity of the translation results of all repeated sentences is maintained; this process may be accomplished by computer automated replacement.
In a second aspect of the present invention, there is provided a translation quality pre-warning control system, which can at least implement the steps of the foregoing method, the system further comprising:
(1) The uploading interface operation piece is configured to upload the current sentence to be translated and the current translation sentence;
as a further improvement of the present invention, the upload interface operation member may be further configured to include:
the automatic uploading setting interface is used for configuring related interface parameters of the automatic uploading setting interface, and when the related interface parameters are met, the current sentence to be translated and the current translation sentence can be automatically uploaded to a corpus database;
the relevant interface parameters include: sentence length of the current sentence to be translated; and/or sentence length of the current translation sentence; the frequency of occurrence of keywords/words of the current sentence to be translated and/or the frequency of occurrence of keywords/words of the current sentence to be translated; whether the current sentence to be translated contains the vocabulary in the concerned database or not and/or whether the current sentence to be translated contains the vocabulary in the concerned database or not.
In particular implementations, parameters that may be set on the automatic upload settings interface include, for example: sentence length threshold or threshold range; for example, setting a threshold range of the length of the current sentence to be translated/the current translation sentence, and when the length of the current sentence to be translated/the current translation sentence meets the threshold range condition, automatically uploading;
similarly, a threshold value of the occurrence frequency/frequency range of the keywords/words can be set, and when the frequency or the frequency of the corresponding keywords/words contained in the current sentence to be translated/the current translation sentence meets the threshold value condition, the keywords/words are automatically uploaded;
similarly, judging parameters can be set to judge whether the current sentence to be translated/the current translation sentence contains the vocabulary in the concerned database or not, and the like;
one of the above parameter settings may be satisfied, or a combination thereof may be used.
The attention database refers to a preset database containing special words needing attention. The translation materials in different fields include different specific words that need to be of interest, for example, words that need to be of interest in firearm translation include Magazine. This may be provided by a translation expert in the relevant field or preset in advance.
If the automatic uploading interface is not configured, manually determining whether to upload the current sentence to be translated and the current translation sentence by a current translator; and/or when the current sentence to be translated and the current translation sentence are not automatically uploaded to the corpus database, the current translator can also manually decide whether to upload the current sentence to be translated and the current translation sentence.
The corpus may be a local database or a remote database.
If a translator translates independently or a plurality of team members in the same network range (such as company intranet) translate together, a local database can be used;
if the translation is performed cooperatively by a plurality of team persons in different network ranges, a shared remote/cloud server can be adopted, and a shared remote database is arranged in the remote/cloud server;
(2) The operation member is repeatedly detected. The repeated detection operation part detects whether repeated sentences to be translated exist in the corpus aiming at the uploaded current sentences to be translated;
and/or;
and detecting whether repeated translation sentences exist in the corpus aiming at the uploaded current translation sentences.
The repeated detection operation part detects whether the repeated operation exists, and can be performed based on the similarity comparison of the current translation sentence/the sentence to be translated and the sentences of the corpus, and can also be performed based on the keyword comparison contained in each; the invention is not limited in this regard.
(3) Quality early warning prompt.
And if repeated sentences to be translated and/or repeated translated sentences are detected, the quality early warning prompt performs early warning on the current translation quality.
The quality early warning prompt part at least adopts one of the following measures to early warn the current translation quality:
i. prompting the current translator: repeated sentences exist in the corpus, and the translation result of the repeated sentences is inconsistent with the translation result given by the current translator;
prompting other translators: a certain translator gives different translation results for the same sentence;
global hint: prompting all translators currently processing the same translation corpus, and different translators of a sentence give different results.
Searching repeated sentences to be translated existing in the remaining untranslated parts of the corpus to be translated currently, and making a marking prompt.
The prompt may include various manners such as different colors, an audio alarm, a pop-up window, etc., and the present invention is not particularly limited.
(4) And a translation result correction piece.
And if the translation quality early warning prompt given by the quality early warning prompt piece is received, the translation result correcting piece corrects the current translation result according to the actual situation. The purpose of the correction is to ensure that the translation results of all repeated sentences are consistent.
Since the early warning prompt shows that different translation results exist, the different translation results must be corrected to a consistent unified result according to a certain standard.
As another innovative point of the present invention, the translation result correction unit includes at least one of the following components:
automatic correction component: and automatically completing correction by adopting an artificial intelligence system. The artificial intelligence system automatically selects an optimal translation result as a consistent unified result according to machine learning, and comprises the following steps: automatically counting translation results of the same sentence in the historical translation corpus, and selecting the translation result with the largest selected number as a uniform result;
cross-alignment correction component: displaying all the translation results aiming at the repeated sentences at present, allowing all the current translators to cross-compare and vote, and selecting the translation result with the largest voting number as a uniform result;
expert correction component: the expert provides the translation result of the current repeated sentence as a consistent unified result.
In expert correction, the real-time online correction of the expert can be realized, and the current sentences to be corrected can be submitted, and unified correction can be performed when a certain number of sentences to be corrected are reached or the expert is online.
In the technical scheme of the invention, the translation result correcting parts can be mutually matched for use, or one of the translation result correcting parts can be selected; it is also possible to set in advance which correction component is to be used preferentially; different correction priorities can be set in advance, and when correction results given by different correction components are inconsistent, the result given by the correction component with the highest priority is selected; when the correction results given by at least two different correction components are identical, the identical correction results are directly employed.
Through the mode, the translation quality can be timely found and early-warned, and quality correction is completed in real time.
Of course, the present invention also discloses a computer readable storage medium, on which computer executable instructions are stored, and when the instructions are executed by a memory and a processor, the steps of the above disclosed method can be completed, and the technical problem of the present application can be solved.
The technical proposal of the invention has the advantages that:
(1) The translator only needs to pay attention to the current translation work, and does not need to worry about whether the translation result of the translator is consistent with other people;
(2) The technical scheme of the invention can automatically detect repeated sentences and judge whether translation quality problems exist or not so as to provide early warning prompt; after the early warning prompt is sent, a corresponding correction result can be given, and the uniformity of the translation results of all repeated sentences is maintained; this process may be accomplished by computer automated replacement.
(3) The translation quality early warning process does not influence the normal translation work, and a translator is not required to stop the current translation work.
More specific embodiments of the present invention will be further presented in the detailed description.
Drawings
FIG. 1 is a system interface frame diagram of a translation quality warning control system of the present invention
FIG. 2 is a flow chart of a translation quality early warning method of the present invention
Detailed Description
Referring to fig. 1, the translation quality early warning control system of the present invention includes an upload interface operation member, a repeated detection operation member, a quality early warning prompt member, and a translation result correction member;
the upload interface manipulation member may be further configured to include: the automatic uploading setting interface is used for configuring related interface parameters of the automatic uploading setting interface, and when the related interface parameters are met, the current sentence to be translated and the current translation sentence can be automatically uploaded to a corpus database;
the relevant interface parameters include: sentence length of the current sentence to be translated; and/or sentence length of the current translation sentence; the frequency of occurrence of keywords/words of the current sentence to be translated and/or the frequency of occurrence of keywords/words of the current sentence to be translated; whether the current sentence to be translated contains the vocabulary in the concerned database or not and/or whether the current sentence to be translated contains the vocabulary in the concerned database or not.
The attention database refers to a preset database containing special words needing attention. The translation materials in different fields include different specific words that need to be of interest, for example, words that need to be of interest in firearm translation include Magazine. This may be provided by a translation expert in the relevant field or preset in advance.
If the automatic uploading interface is not configured, manually determining whether to upload the current sentence to be translated and the current translation sentence by a current translator; and/or when the current sentence to be translated and the current translation sentence are not automatically uploaded to the corpus database, the current translator can also manually decide whether to upload the current sentence to be translated and the current translation sentence.
The prompt of the quality early warning prompt piece can comprise various modes such as different colors, sound alarms, popup windows and the like, and the invention is not particularly limited.
The translation result correction part at least comprises one of the following components:
automatic correction component: and automatically completing correction by adopting an artificial intelligence system. The artificial intelligence system automatically selects an optimal translation result as a consistent unified result according to machine learning, and comprises the following steps: automatically counting translation results of the same sentence in the historical translation corpus, and selecting the translation result with the largest selected number as a uniform result;
cross-alignment correction component: displaying all the translation results aiming at the repeated sentences at present, allowing all the current translators to cross-compare and vote, and selecting the translation result with the largest voting number as a uniform result;
expert correction component: the expert provides the translation result of the current repeated sentence as a consistent unified result.
In fig. 1, the translation result correction includes three components, which are shown separately for convenience of illustration.
Referring to fig. 2, the translation quality early warning method of the present invention includes the following four steps:
(1) Uploading a current sentence to be translated and a current translation sentence;
(2) Detecting repeated sentences to be translated and/or repeated translated sentences;
(3) Translation quality early warning prompt;
(4) And correcting the translation result.
In this embodiment, the current sentence to be translated refers to a sentence to be translated currently by a translator, and the current translated sentence refers to a translation result given by the translator for the current sentence to be translated;
in this embodiment, the current sentence to be translated and the current translated sentence may be uploaded to a corpus, where the corpus may be a local database or a remote database. If a translator translates independently or a plurality of team members in the same network range (such as company intranet) translate together, a local database can be used; if the translation is performed cooperatively by a plurality of team persons in different network ranges, a shared remote/cloud server can be adopted, and a shared remote database is arranged in the remote/cloud server;
in this embodiment, the repeated sentences to be translated refers to sentences whose similarity with the current sentences to be translated uploaded by the current translator in the corpus meets a first condition;
in this embodiment, the repeated translation sentence refers to a sentence that exists in the corpus and has similarity with the current translation sentence uploaded by the current translator and satisfies the second condition.
The inventors found that the translation process had the following phenomena:
a) Different sentences, the translation results may be the same;
b) The translation results may be different for the same sentence.
However, in any case, there is a certain similarity between different translation results of these different sentences. When the similarity satisfies a certain condition, it means that these different sentences/different translation results should be kept uniform.
In this embodiment, once the repeated sentence to be translated and/or the repeated translated sentence is detected, at least one of the following measures is taken:
i. prompting the current translator: repeated sentences exist in the corpus, and the translation result of the repeated sentences is inconsistent with the translation result given by the current translator;
prompting other translators: a certain translator gives different translation results for the same sentence;
global hint: prompting all translators currently processing the same translation corpus, and different translators of a sentence give different results.
Searching repeated sentences to be translated existing in the remaining untranslated parts of the corpus to be translated currently, and making a marking prompt.
The prompt may include various manners such as different colors, an audio alarm, a pop-up window, etc., and the present invention is not particularly limited.
Through the mode, the translation quality can be timely found and early-warned, and quality correction is completed in real time. The process of quality correction may take many forms, such as the automatic correction mentioned previously, cross-over correction, expert correction.
In this embodiment, the translation result quality correction methods of the above multiple methods may be used in combination with each other, or one of them may be selected; it is also possible to set in advance which correction mode is preferred; the priorities of different correction modes can be set in advance, and when the correction results given by the different correction modes are inconsistent, the result given by the correction mode with the highest priority is selected; when the correction results given by at least two different correction modes are identical, the identical correction results are directly adopted.
Claims (7)
1. A translation quality early warning control system is characterized in that:
the system comprises the following configuration parts:
the uploading interface operation piece is configured to upload the current sentence to be translated and the current translation sentence to the corpus;
the uploading interface operation piece is configured to comprise an automatic uploading setting interface, and the current sentence to be translated and the current translation sentence can be automatically uploaded to a corpus database by configuring relevant interface parameters of the automatic uploading setting interface when the relevant interface parameters are met;
the relevant interface parameters include: sentence length of the current sentence to be translated; and/or sentence length of the current translation sentence; the frequency of occurrence of keywords/words of the current sentence to be translated and/or the frequency of occurrence of keywords/words of the current sentence to be translated; whether the current sentence to be translated contains the vocabulary in the concerned database or not, and/or whether the current sentence to be translated contains the vocabulary in the concerned database or not;
the concerned database is a preset database containing special words needing to be concerned;
the repeated detection operation part is used for detecting whether repeated sentences to be translated exist in the corpus aiming at the uploaded current sentences to be translated; and/or, for the uploaded current translation sentence, detecting whether repeated translation sentences exist in the corpus;
quality early warning prompting piece: if repeated sentences to be translated and/or repeated translated sentences are detected, the quality early warning prompt performs early warning on the current translation quality;
translation result correction piece: and if the translation quality early warning prompt given by the quality early warning prompt piece is received, the translation result correcting piece corrects the current translation result according to the actual situation.
2. The translation quality warning control system according to claim 1, wherein: if the automatic uploading setting interface is not configured, manually determining whether to upload the current sentence to be translated and the current translation sentence by a current translator; and/or when the current sentence to be translated and the current translation sentence are not automatically uploaded to the corpus database, manually determining whether to upload the current sentence to be translated and the current translation sentence by a current translator.
3. The translation quality warning control system according to claim 1 or 2, characterized in that: the translation result correction part at least comprises one of the following components:
automatic correction component: automatically counting translation results of the same sentence in the historical translation corpus, and selecting the translation result with the largest selected number as a uniform result;
cross-over alignment correction component: displaying all the translation results aiming at the repeated sentences at present, allowing all the current translators to cross-compare and vote, and selecting the translation result with the largest voting number as a uniform result; expert correction component: the expert provides the translation result of the current repeated sentence as a consistent unified result.
4. A translation quality early warning method based on repeated sentence detection, the method being performed by a translation quality early warning control system according to any one of claims 1 to 3, comprising the steps of:
(1) Uploading a current sentence to be translated and a current translation sentence;
(2) Repeating the steps of sentence waiting and/or sentence translation detection;
(3) A translation quality early warning step;
(4) Correcting translation results;
the method is characterized in that:
before the step (1), a judging step is further included, wherein the judging step is used for judging whether to execute the uploading step;
and, upon detecting a duplicate to-be-translated sentence and/or a duplicate translated sentence, taking at least one of the following actions:
prompting a current translator, prompting other translators, globally prompting, searching repeated sentences to be translated which exist in the residual untranslated part of the current corpus to be translated, and marking the prompt.
5. The translation quality early warning method based on repeated sentence detection as claimed in claim 4, wherein: the step (4) comprises: the correction is performed in one or more of the following ways:
(81) Automatic correction: automatically completing correction by adopting an artificial intelligent system;
(82) Cross comparison correction: displaying all the translation results aiming at the repeated sentences at present, allowing all the current translators to cross-compare and vote, and selecting the translation result with the largest voting number as a uniform result;
(83) Expert correction: the expert provides the translation result of the current repeated sentence as a consistent unified result.
6. The translation quality early warning method based on repeated sentence detection as claimed in claim 5, wherein: setting priorities of different correction modes in advance, and selecting a result given by a correction mode with the highest priority when correction results given by different correction modes are inconsistent; when the correction results given by at least two different correction modes are identical, the identical correction results are directly adopted.
7. A computer readable storage medium having stored thereon computer readable instructions for execution by a memory and a processor for performing the method of any of claims 4-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811194428.9A CN109408832B (en) | 2018-10-16 | 2018-10-16 | Translation quality early warning method and system based on repeated sentence detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811194428.9A CN109408832B (en) | 2018-10-16 | 2018-10-16 | Translation quality early warning method and system based on repeated sentence detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109408832A CN109408832A (en) | 2019-03-01 |
| CN109408832B true CN109408832B (en) | 2023-06-16 |
Family
ID=65467872
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811194428.9A Active CN109408832B (en) | 2018-10-16 | 2018-10-16 | Translation quality early warning method and system based on repeated sentence detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109408832B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111274809A (en) * | 2020-02-20 | 2020-06-12 | 苏宁云计算有限公司 | Method and device for processing linguistic data in knowledge base |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI386822B (en) * | 2007-09-05 | 2013-02-21 | Shing Lung Chen | A method for establishing a multilingual translation data base rapidly |
| US20090192782A1 (en) * | 2008-01-28 | 2009-07-30 | William Drewes | Method for increasing the accuracy of statistical machine translation (SMT) |
| CN103678285A (en) * | 2012-08-31 | 2014-03-26 | 富士通株式会社 | Machine translation method and machine translation system |
| CN105183723A (en) * | 2015-09-17 | 2015-12-23 | 成都优译信息技术有限公司 | Associating method for translation software and language material searching |
-
2018
- 2018-10-16 CN CN201811194428.9A patent/CN109408832B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109408832A (en) | 2019-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6271085B2 (en) | Learning system, learning device, learning method, learning program, teacher data creation device, teacher data creation method, teacher data creation program, terminal device, and threshold change device | |
| CN111309912B (en) | Text classification method, apparatus, computer device and storage medium | |
| EP3407200B1 (en) | Method and device for updating online self-learning event detection model | |
| US10607463B2 (en) | Automated object and activity tracking in a live video feed | |
| CN105930432B (en) | Training method and device for sequence labeling tool | |
| JP2015087903A (en) | Apparatus and method for information processing | |
| WO2018024243A1 (en) | Method and device for verifying recognition result in character recognition | |
| WO2020244183A1 (en) | Data annotation | |
| US20190034807A1 (en) | Generating a consistently labeled training dataset by automatically generating and displaying a set of most similar previously-labeled texts and their previously assigned labels for each text that is being labeled for the training dataset | |
| EP3929800A1 (en) | Skill word evaluation method and device, electronic device, and computer readable medium | |
| CN106610990B (en) | Method and device for analyzing emotional tendency | |
| CN111310440A (en) | Text error correction method, device and system | |
| CN105843889A (en) | Credibility based big data and general data oriented data collection method and system | |
| CN109408832B (en) | Translation quality early warning method and system based on repeated sentence detection | |
| US10885593B2 (en) | Hybrid classification system | |
| JP6988995B2 (en) | Image generator, image generator and image generator | |
| CN112199805B (en) | Power transmission line hidden danger identification model evaluation method and device | |
| CN111339285B (en) | BP neural network-based enterprise resume screening method and system | |
| CN115827899A (en) | Data integration method, device and equipment based on knowledge graph and storage medium | |
| CN113554062B (en) | Training method, device and storage medium for multi-classification model | |
| CN113777965B (en) | Spray quality control method, spray quality control device, computer equipment and storage medium | |
| CN115169328A (en) | High-accuracy Chinese spelling check method, system and medium | |
| CN109359306B (en) | Translation correction method and system based on repeated sentence detection | |
| JP2021144619A (en) | Information processing device, information processing method and computer program | |
| US20190197172A1 (en) | System and method for data analysis and presentation of data |
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 |