CN107273363B - A kind of language text interpretation method and system - Google Patents
A kind of language text interpretation method and system Download PDFInfo
- Publication number
- CN107273363B CN107273363B CN201710335652.4A CN201710335652A CN107273363B CN 107273363 B CN107273363 B CN 107273363B CN 201710335652 A CN201710335652 A CN 201710335652A CN 107273363 B CN107273363 B CN 107273363B
- Authority
- CN
- China
- Prior art keywords
- translation
- probability distribution
- text
- probability
- language text
- 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
- 238000000034 method Methods 0.000 title claims abstract description 29
- 241000208340 Araliaceae Species 0.000 claims 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims 2
- 235000008434 ginseng Nutrition 0.000 claims 2
- 238000013528 artificial neural network Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 230000007306 turnover Effects 0.000 description 4
- 230000001537 neural effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
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
- G06F40/42—Data-driven translation
- G06F40/44—Statistical methods, e.g. probability models
-
- 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
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 present invention provides a kind of language text interpretation method and system.This method comprises: determining rule according to preset translation candidate collection, determine that the corresponding translation candidate collection of source language text, the translation candidate collection include multiple cypher texts of source language text;The source language text is language text to be translated;Based on the translation candidate collection, preset translation model and preset priori knowledge model, the first probability distribution and the second probability distribution are determined;First probability distribution is used to indicate the probability that the cypher text meets priori knowledge model, and second probability distribution is used to indicate the probability that the cypher text meets translation model;Based on first probability distribution and second probability distribution, the cypher text of the source language text is determined from the translation candidate collection.The present invention can incorporate any priori knowledge in translation model, to improve the accuracy and reliability of machine translation.
Description
Technical field
The present invention relates to machine translation mothod field, in particular to a kind of language text interpretation method and system.
Background technique
With international progress, the exchange between different language crowd is growing day by day, translate into order in exchanging to closing
Important tool.Machine translation because it is convenient simple and free the advantages that, greatly meet the translation demand of people, improve
The efficiency of international exchange, so that more stringent requirements are proposed for correctness of the people to machine translation.
Machine translation can substantially be divided into: rule-based machine translation method and the machine translation based on corpus.Base
In the machine translation of corpus, its critical issue, which is that, establishes a complete corpus, alternatively referred to as high quality
Training sample.The training sample of high quality directly affects the accuracy of translation.However, establishing the training sample of high quality not
It is an easy thing, reason is that sample data is limited, and cannot portray the distribution of initial data well;In addition,
Even if sample data is enough, it can not avoid wherein the presence of error sample, i.e. noise data.The mind obtained based on the training sample
It is difficult to prepare to embody master mould through network, or even will appear the case where violating priori knowledge.In this case, priori knowledge
Introducing just becomes particularly significant.For translation rule, for example, " should not repeat translation, should not also leak and turn over ", such rule is just
It can be described as priori knowledge.Many studies have shown that incorporating priori knowledge in neural network model to constrain it, mind can be improved
Performance through network.
Machine translation method (the Attention-based Neural Machine of neural network based on attention mechanism
Translation;Abbreviation Attention-based NMT) be the machine translation based on corpus a branch, and at present
A kind of machine translation method used in mainstream translation system.Its basic thought is using a non-linear neural net end to end
Source language text is directly mapped to target language text by network, that is, constructs the new frame of one " coding-decoding ": giving a source
Language sentence is mapped as a continuous, dense vector using an encoder first, then reuses a decoder
A target language sentence is converted by the vector.But this method is difficult for priori knowledge to be dissolved among neural network.
There are also the technologies being dissolved into priori knowledge in neural network at present.For example, some technologies are by priori knowledge
It is indicated with additional neural network module;Some technologies are by adding limit entry in training objective to incorporate priori knowledge.Though
These right technologies can promote translation effect significantly, but the former require the correlation between different priori knowledges be also required to by
Modeling, the latter are merely able to add a small amount of simple limit entry.These problems cause these technologies that cannot be applied to will be any, multiple
Miscellaneous priori knowledge incorporates neural network machine translation model.
Therefore, how a kind of interpretation method that any priori knowledge can be incorporated to neural network machine translation model is provided
It is a urgent problem needed to be solved.
Summary of the invention
For solve the problems, such as it is of the existing technology can not by any priori knowledge incorporate neural network translation model, this hair
It is bright that a kind of language text interpretation method and system are provided.
On the one hand, the present invention provides a kind of language text interpretation method, this method comprises:
Rule is determined according to preset translation candidate collection, determines the corresponding translation candidate collection of source language text, it is described
Translation candidate collection includes multiple cypher texts of source language text;The source language text is language text to be translated;
Based on the translation candidate collection, preset translation model and preset priori knowledge model, the first probability is determined
Distribution and the second probability distribution;First probability distribution is used to indicate the cypher text and meets the general of priori knowledge model
Rate, second probability distribution are used to indicate the probability that the cypher text meets translation model;
Based on first probability distribution and second probability distribution, the source is determined from the translation candidate collection
The cypher text of language text.
On the other hand, the present invention provides a kind of language text translation system, which includes:
Candidate collection module is translated, for determining rule according to preset translation candidate collection, determines source language text pair
The translation candidate collection answered, the translation candidate collection include multiple cypher texts of source language text;The source language text
For language text to be translated;
Training module, for being based on the translation candidate collection, preset translation model and preset priori knowledge model,
Determine the first probability distribution and the second probability distribution;First probability distribution, which is used to indicate the cypher text and meets priori, to be known
Know the probability of model, second probability distribution is used to indicate the probability that the cypher text meets translation model;
Translation module, for being based on first probability distribution and second probability distribution, from the translation Candidate Set
The cypher text of the source language text is determined in conjunction.
Language text interpretation method provided by the invention and system, by calculating separately priori knowledge model and translation model
Probability distribution in translation candidate collection, and using the difference of two probability distribution as a part of speech training target, from
And Machine Translation Model is made to may learn arbitrary priori knowledge, improve the accuracy of machine translation result and reliable
Property.
Detailed description of the invention
Fig. 1 is the flow diagram of language text interpretation method provided in an embodiment of the present invention;
Fig. 2 is the structural schematic diagram of language text translation system provided in an embodiment of the present invention;
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical solution in the embodiment of the present invention is explicitly described, it is clear that described embodiment is the present invention
A part of the embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not having
Every other embodiment obtained under the premise of creative work is made, shall fall within the protection scope of the present invention.
Fig. 1 is the flow diagram of language text interpretation method provided in an embodiment of the present invention.As shown in Figure 1, this method
The following steps are included:
Step 101 determines rule according to preset translation candidate collection, determines the corresponding translation Candidate Set of source language text
It closes, the translation candidate collection includes multiple cypher texts of source language text;The source language text is language to be translated
Text;
Step 102 is based on the translation candidate collection, preset translation model and preset priori knowledge model, determines
First probability distribution and the second probability distribution;First probability distribution is used to indicate the cypher text and meets priori knowledge mould
The probability of type, second probability distribution are used to indicate the probability that the cypher text meets translation model;
Step 103 is based on first probability distribution and second probability distribution, from the translation candidate collection really
The cypher text of the fixed source language text.
Specifically, firstly, preset translation candidate collection determines that rule refers to that translation is the task that a sequence generates, source
There are multiple words or word in language text x, when generating translation candidate collection, the word or word of previous generation can be used as latter
The input of a word or word.According to the source language text x of different length, it is exponential for really translating the size of candidate collection
, it can not effectively calculate.In practical applications, by stochastical sampling or beam search, to obtain the source language text
Multiple cypher texts, i.e., translation candidate collection S (x), can be realized using the prior art, details are not described herein again;
Then, according to the translation candidate collection S (x) and preset priori knowledge model Q (y | x;γ), the first probability is determined
DistributionAccording to the translation candidate collection S (x) and preset translation model P (y | x;θ), the second probability distribution is determinedFinally, being based on the first probability distribution and the second probability distribution, source language text is determined from translation candidate collection
Cypher text y.
For sake of clarity, if source language text x, as input, cypher text y thus constitutes sentence pair as output
(x, y).In practical applications, under different contexts the same word or word there are different semantemes, and source language text x be by
Multiple words or word are according to the different compositions that puts in order, and the uncertainty of the ambiguity and sequence of word or word leads to one
A source language text may correspond to multiple cypher texts (y1, y2, y3 etc.), and probability is highest in this multiple cypher text, be
Best cypher text, in order to be distinguished with other cypher texts, referred to as target language text.
For example, preset priori knowledge model Q (y | x;γ), it can be obtained not according to different characteristic function φ (x, y)
Same model, the first probability distribution can determine according to the following formula:
Wherein, x indicates source language text, and y is target language text, and y ' is cypher text, and γ is priori knowledge model
Parameter preset.
Characteristic function φ (x, y) indicates the corresponding relationship of source language text and cypher text in priori knowledge data base,
It based on specific characteristic function, is given a mark, that is, calculated each to each cypher text y1, y2 and y3 using priori knowledge model
Cypher text meets the probability of priori knowledge model.Wherein, more meet the cypher text of priori knowledge model, probability is higher.
Translation model P (y | x;It θ) is then the commonly used scoring model of machine translation, which can be parallel by training
Corpus obtains, and the corresponding relationship of source language text x and cypher text y in Parallel Corpus is indicated, for calculating each translation
Text meets the probability of translation model, belongs to the prior art, and details are not described herein again.
According to translation candidate collection S (x) and translation model P (y | x;θ), the second probability distribution can be determined by following formula:
Wherein, x indicates source language text, and y is target language text, and y ' is cypher text, and θ is the parameter of translation model;α
It is the default hyper parameter for controlling the second probability distribution steep.
Language text interpretation method provided in an embodiment of the present invention passes through comprehensive utilization priori knowledge model and translation mould
Type gives a mark to multiple cypher texts in terms of two, so that the cypher text for more meeting priori knowledge model be encouraged to turn over
The probability translated under model is also higher, to finally determine target language text from translation candidate collection, improves translation model
Performance and translation result accuracy.
On the basis of the above embodiments, first probability distribution and described second in the language text interpretation method
Probability distribution determines the cypher text of the source language text from the translation candidate collection, comprising:
Based on first probability distribution and second probability distribution, probability difference parameter value is determined;The probability difference
Different parameter is used to indicate the difference of first probability distribution and second probability distribution;
Based on the probability difference parameter value, the translation text of the source language text is determined from the translation candidate collection
This.
Specifically, firstly, determining rule according to preset translation candidate collection, the corresponding translation of source language text x is determined
Candidate collection S (x);Then, it is based on the translation candidate collection, translation model and priori knowledge model, determines the first probability distributionWith the second probability distributionLater, it determines general between the first probability distribution and the second probability distribution
Rate difference parameter value;Finally, being based on the probability difference parameter value, determine source language text x's from translation candidate collection S (x)
Cypher text y.
For example, user log in translation system after, in-English translation window input in Chinese column in input source language text x
For " many airports are all forced to close ", determining translation candidate collection S (x) according to x, there are two cypher texts: y1 is " Many
Airports were closed to close " and y2 is " Many airports were forced to close
down";
According to priori knowledge model, the first probability distribution is determined
Wherein, the probability that Q (y1 | x)=0.2, i.e. sentence pair (x, y1) meet priori knowledge model is 0.2;Q (y2 | x)=
0.8, i.e., it is 0.8 that sentence pair (x, y2), which meets the probability of priori knowledge model,;
According to translation model, the second probability distribution is determined:
Wherein, the probability that P (y1 | x)=0.6, i.e. sentence pair (x, y1) meet translation model is 0.6;P (y2 | x)=0.4, i.e.,
The probability that sentence pair (x, y2) meets translation model is 0.4;
By the first probability distribution and the second probability distribution, difference parameter value between the two can be determined;Based on the difference
Different parameter value is adjusted translation model and gives a mark again to above-mentioned two cypher text, obtain P (y1 | x)=0.3, P (y2 |
X)=0.7;
Accordingly, it is determined that source language text x: the cypher text y: " Many airports of " many airports are all forced to close "
were forced to close down”。
By above-described embodiment, it can be seen that, language text interpretation method provided in an embodiment of the present invention is based on the first probability
The difference parameter value of distribution and the second probability distribution, and given a mark again according to translation model to multiple cypher texts, to improve
Meet probability of the cypher text of priori knowledge in translation model probability distribution, and then obtains more accurate source language text
Cypher text.
On the basis of the above embodiments, the difference parameter value of first probability distribution and second probability distribution is
KL (Kullback-Leibler) distance can be determined by following formula:
On the basis of the various embodiments described above, in the language text interpretation method based on the probability difference parameter value,
The cypher text of the source language text is determined from the translation candidate collection, comprising:
Based on the difference parameter value, training objective is determined;The training objective is used to indicate the translation model to institute
State priori knowledge Model approximation;
Based on the training objective and the preset model that reorders, the original language is determined from the translation candidate collection
The cypher text of text.
Specifically, firstly, determining rule according to preset translation candidate collection, the corresponding translation of source language text x is determined
Candidate collection S (x);Then, it is based on the translation candidate collection, translation model and priori knowledge model, determines the first probability distributionWith the second probability distributionLater, it determines general between the first probability distribution and the second probability distribution
Rate difference parameter value;Finally, being based on the probability difference parameter value, training objective J (θ, γ) is determined, so that translation model is to priori
Model approximation;Finally, being based on training objective J (θ, γ) and the preset model that reorders, determined from translation candidate collection S (x)
The cypher text y of source language text x.
In general, when giving a mark to cypher text, generally use translation model P (y | x;Log-likelihood θ) is estimated
Be counted as standard exercise criterion, i.e., traditional training objective be log-likelihood function L (θ)=logP (y | x;θ).
By determining the difference parameter value of the first probability distribution and the second probability distribution, which is added tradition
In training objective, determine that new training objective is J (θ, γ), which thinks that optimal parameter θ and γ can encourage most to accord with
Probability highest of the cypher text of priori knowledge in the second probability distribution of translation model is closed, so that translation model more inclines
The cypher text that priori knowledge is determined for compliance in Xiang Yucong translation candidate collection S (x) is the target language text of source language text x
y。
Optionally, if the difference parameter value is KL distance, training objective can determine according to the following formula:
Wherein, λ1And λ2It is the default hyper parameter of balance training target, N is the sentence pair number of training data.
Optimal parameter θ and γ is obtained by new training objective, using the following model that reorders, from translation candidate
Determine the cypher text of source language text.
Y=argmaxy∈S(x){logP(y|x;θ)+γ·φ(x,y)}
For example, it is assumed that source language text x is " Bush and salon have held talks ", translation candidate collection S is determined according to x
(x) there are three cypher texts: y1 is " Bush held a talk with Sharon ", and y2 is " Bush held a talk
With Bush ", y3 are " Bush had lunch with Sharon ".
Assuming that characteristic function φ (x, y) indicates the word pair occurred in source language text x and target language text y in sentence pair
Quantity, word is combined into { (Bush, Bush), (holding, held), (talks, talk), (salon, Sharon) } to collection, then the
In one cypher text y1,4 words are to occurring, therefore φ (x, y1)=4;Similarly, φ (x, y2)=3, φ (x, y3)=
2。
The first probability distribution can be determined according to priori knowledge model
Wherein, the probability of cypher text y1 are as follows:
It can similarly obtain: Q (y2 | x)=e3/(e2+e3+e4);Q (y3 | x)=e2/(e2+e3+e4).Final Q (y1 | x)=
0.67, Q (y2 | x)=0.24, Q (y3 | x)=0.09.
It by above-mentioned probability it is found that cypher text y1 is best suitable for priori knowledge model, and is in fact also correctly to turn over
Translation sheet;Cypher text y2 has then obviously violated the priori knowledge of " should not repeat translation, should not leak and turn over ", therefore probability is lower;
Cypher text y3 then deviates from the semanteme of source language text, therefore probability is lower.
Assuming that obtaining the second probability distribution by adjusting preceding translation model
Wherein, P (y1 | x)=0.4, P (y2 | x)=0.5, P (y3 | x)=0.1, translation model can translate " Bush
held a talk with Bush”。
At this point, if default hyper parameter λ1、λ2Numerical value be 1, KL (P between above-mentioned two probability distribution is calculated by formula
| | Q), new training objective J (θ, γ) is determined based on KL distance;
It based on the training objective and reorders model, translation model is adjusted, the P (y1 | x)=0.6 after training,
P (y2 | x)=0.31, P (y3 | x)=0.09, it is seen then that new training objective improves the probability of cypher text y1, and reduces
The probability of cypher text y2 and y3, so that more meeting the cypher text of priori knowledge probability in the probability distribution in translation model
It is higher, even if translation model is to priori knowledge Model approximation.
Therefore, the target language text y of final output is " Bush held a talk with Sharon ".
By above-described embodiment it can be seen that, language text interpretation method provided in an embodiment of the present invention, by the way that elder generation will be met
Test the probability distribution of knowledge model and meet translation model probability distribution between KL distance traditional training objective, drum is added
It encourages more meeting the translation that is also higher, and then more being optimized of probability of the cypher text of priori knowledge model under translation model
Model parameter improves the performance and translation of translation model to finally determine target language text from translation candidate collection
As a result accuracy.
Fig. 2 is the structural schematic diagram of language text translation system provided in an embodiment of the present invention.As shown in Fig. 2, the system
It include: translation candidate collection module 21, training module 22 and translation module 23.Wherein, translation candidate collection module 21 is used for root
Rule is determined according to preset translation candidate collection, determines the corresponding translation candidate collection of source language text, the translation Candidate Set
Close multiple cypher texts including source language text;The source language text is language text to be translated;Training module 22 is used
In be based on the translation candidate collection, preset translation model and preset priori knowledge model, determine the first probability distribution and
Second probability distribution;First probability distribution is used to indicate the probability that the cypher text meets priori knowledge model, described
Second probability distribution is used to indicate the probability that the cypher text meets translation model;Translation module 23 is used to be based on described first
Probability distribution and second probability distribution determine the cypher text of the source language text from the translation candidate collection.
It should be noted that the language text translation system is to realize that above method embodiment, function are specific
It can refer to above method embodiment, details are not described herein again.
On the basis of the above embodiments, the translation module 23 in the system is specifically used for being based on first probability distribution
And second probability distribution, determine probability difference parameter value;The probability difference parameter is used to indicate first probability point
The difference of cloth and second probability distribution;Based on the probability difference parameter value, institute is determined from the translation candidate collection
State the cypher text of source language text.Optionally, the probability difference parameter is KL distance.
On the basis of the various embodiments described above, the translation module 23 in the system is specifically used for being based on the difference parameter
Value, determines training objective;The training objective is used to indicate the translation model to the priori knowledge Model approximation;Based on institute
Training objective and the preset model that reorders are stated, the translation text of the source language text is determined from the translation candidate collection
This.
Priori knowledge is dissolved into translation in the training stage by the language text interpretation method and system provided through the invention
In model, the performance of translation model is improved, and then priori knowledge is applied in translation process, without increasing additionally
Network module, which achieves that, applies to any priori knowledge in machine translation, the final accuracy for improving translation result and reliable
Property.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features;
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (4)
1. a kind of language text interpretation method characterized by comprising
Rule is determined according to preset translation candidate collection, determines the corresponding translation candidate collection of source language text, the translation
Candidate collection includes multiple cypher texts of source language text;The source language text is language text to be translated;
Based on the translation candidate collection, preset translation model and preset priori knowledge model, the first probability distribution is determined
And second probability distribution;First probability distribution is used to indicate the probability that the cypher text meets priori knowledge model, institute
It states the second probability distribution and is used to indicate the probability that the cypher text meets translation model;
Based on first probability distribution and second probability distribution, the original language is determined from the translation candidate collection
The cypher text of text;
It is described to be based on first probability distribution and second probability distribution, the source is determined from the translation candidate collection
The cypher text of language text, comprising:
Based on first probability distribution and second probability distribution, probability difference parameter value is determined;The probability difference ginseng
Number is used to indicate the difference of first probability distribution and second probability distribution;
Based on the probability difference parameter value, the cypher text of the source language text is determined from the translation candidate collection;
It is described to be based on the probability difference parameter value, the translation text of the source language text is determined from the translation candidate collection
This, comprising:
Based on the difference parameter value, training objective is determined;The training objective is used to indicate the translation model to the elder generation
Knowledge model is tested to approach;
Based on the training objective and the preset model that reorders, the source language text is determined from the translation candidate collection
Cypher text.
2. the method according to claim 1, wherein the probability difference parameter is KL distance.
3. a kind of language text translation system characterized by comprising
Candidate collection module is translated, for determining rule according to preset translation candidate collection, determines that source language text is corresponding
Candidate collection is translated, the translation candidate collection includes multiple cypher texts of source language text;The source language text be to
The language text of translation;
Training module is determined for being based on the translation candidate collection, preset translation model and preset priori knowledge model
First probability distribution and the second probability distribution;First probability distribution is used to indicate the cypher text and meets priori knowledge mould
The probability of type, second probability distribution are used to indicate the probability that the cypher text meets translation model;
Translation module, for being based on first probability distribution and second probability distribution, from the translation candidate collection
Determine the cypher text of the source language text;
The translation module is specifically used for:
Based on first probability distribution and second probability distribution, probability difference parameter value is determined;The probability difference ginseng
Number is used to indicate the difference of first probability distribution and second probability distribution;
Based on the probability difference parameter value, the cypher text of the source language text is determined from the translation candidate collection;
The translation module is specifically used for:
Based on the difference parameter value, training objective is determined;The training objective is used to indicate the translation model to the elder generation
Knowledge model is tested to approach;
Based on the training objective and the preset model that reorders, the source language text is determined from the translation candidate collection
Cypher text.
4. system according to claim 3, which is characterized in that the probability difference parameter is KL distance.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710335652.4A CN107273363B (en) | 2017-05-12 | 2017-05-12 | A kind of language text interpretation method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710335652.4A CN107273363B (en) | 2017-05-12 | 2017-05-12 | A kind of language text interpretation method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107273363A CN107273363A (en) | 2017-10-20 |
CN107273363B true CN107273363B (en) | 2019-11-22 |
Family
ID=60074224
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710335652.4A Active CN107273363B (en) | 2017-05-12 | 2017-05-12 | A kind of language text interpretation method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107273363B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109783824B (en) * | 2018-12-17 | 2023-04-18 | 北京百度网讯科技有限公司 | Translation method, device and storage medium based on translation model |
CN110298045B (en) * | 2019-05-31 | 2023-03-24 | 北京百度网讯科技有限公司 | Machine translation method, device, equipment and storage medium |
CN110334359B (en) * | 2019-06-05 | 2021-06-15 | 华为技术有限公司 | Text translation method and device |
CN111178085B (en) * | 2019-12-12 | 2020-11-24 | 科大讯飞(苏州)科技有限公司 | Text translator training method, and professional field text semantic parsing method and device |
CN111368091B (en) * | 2020-02-13 | 2023-09-22 | 中国工商银行股份有限公司 | Document translation method and device |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103646019A (en) * | 2013-12-31 | 2014-03-19 | 哈尔滨理工大学 | Method and device for fusing multiple machine translation systems |
CN103678285A (en) * | 2012-08-31 | 2014-03-26 | 富士通株式会社 | Machine translation method and machine translation system |
CN105573994A (en) * | 2016-01-26 | 2016-05-11 | 沈阳雅译网络技术有限公司 | Statistic machine translation system based on syntax framework |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9530161B2 (en) * | 2014-02-28 | 2016-12-27 | Ebay Inc. | Automatic extraction of multilingual dictionary items from non-parallel, multilingual, semi-structured data |
-
2017
- 2017-05-12 CN CN201710335652.4A patent/CN107273363B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103678285A (en) * | 2012-08-31 | 2014-03-26 | 富士通株式会社 | Machine translation method and machine translation system |
CN103646019A (en) * | 2013-12-31 | 2014-03-19 | 哈尔滨理工大学 | Method and device for fusing multiple machine translation systems |
CN105573994A (en) * | 2016-01-26 | 2016-05-11 | 沈阳雅译网络技术有限公司 | Statistic machine translation system based on syntax framework |
Non-Patent Citations (2)
Title |
---|
Knowledge-Based Semantic Embedding for Machine Translation;Chen Shi等;《Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics》;20160812;第2245-2254页 * |
N-Best句法知识增强的统计机器翻译预调序模型;郭俊博 等;《计算机工程与应用》;20160901;第52卷(第17期);第160-165页 * |
Also Published As
Publication number | Publication date |
---|---|
CN107273363A (en) | 2017-10-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107273363B (en) | A kind of language text interpretation method and system | |
JP7122341B2 (en) | Method and apparatus for evaluating translation quality | |
Wu et al. | " Mask and Infill": Applying Masked Language Model to Sentiment Transfer | |
Gu et al. | Dialogwae: Multimodal response generation with conditional wasserstein auto-encoder | |
CN109086357A (en) | Sensibility classification method, device, equipment and medium based on variation autocoder | |
WO2021134520A1 (en) | Voice conversion method, voice conversion training method, intelligent device and storage medium | |
CN106649272A (en) | Named entity recognizing method based on mixed model | |
Tarnavskyi et al. | Ensembling and knowledge distilling of large sequence taggers for grammatical error correction | |
CN116127020A (en) | Method for training generated large language model and searching method based on model | |
CN103646019A (en) | Method and device for fusing multiple machine translation systems | |
CN107861954A (en) | Information output method and device based on artificial intelligence | |
CN103119584A (en) | Machine translation evaluation device and method | |
CN108021549A (en) | Sequence conversion method and device | |
WO2023109294A1 (en) | Method and apparatus for jointly training natural language processing model on basis of privacy protection | |
CN113901208B (en) | Method for analyzing emotion tendentiousness of mid-cross language comments blended with theme characteristics | |
CN105868187A (en) | A multi-translation version parallel corpus establishing method | |
Gang et al. | Chinese intelligent chat robot based on the AIML language | |
Liu et al. | Augmenting multi-turn text-to-SQL datasets with self-play | |
Han et al. | [Retracted] The Modular Design of an English Pronunciation Level Evaluation System Based on Machine Learning | |
CN116610795B (en) | Text retrieval method and device | |
CN109446535A (en) | A kind of illiteracy Chinese nerve machine translation method based on triangle framework | |
Kumar et al. | Towards building text-to-speech systems for the next billion users | |
Iranzo-Sánchez et al. | From simultaneous to streaming machine translation by leveraging streaming history | |
Hao et al. | Abstractive summarization model with a feature-enhanced seq2seq structure | |
CN111724767B (en) | Spoken language understanding method based on Dirichlet variation self-encoder and related equipment |
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 |