43 results about "Machine translated" patented technology

A speech translation system and methods for cross-lingual communication that enable users to improve and modify content and usage of the system and easily abort or reset translation. The system includes a speech recognition module configured for accepting an utterance, a machine translation module, an interface configured to communicate the utterance and proposed translation, a correction module and an abort action unit that removes any hypotheses or partial hypotheses and terminates translation. The system also includes modules for storing favorites, changing language mode, automatically identifying language, providing language drills, viewing third party information relevant to conversation, among other things.

Machine translation decoding is accomplished by receiving as input a text segment in a source language to be translated into a target language, generating an initial translation as a current target language translation, applying one or more modification operators to the current target language translation to generate one or more modified target language translations, determining whether one or more of the modified target language translations represents an improved translation in comparison with the current target language translation, setting a modified target language translation as the current target language translation, and repeating these steps until occurrence of a termination condition. Automatically generating a tree (e.g., either a syntactic tree or a discourse tree) can be accomplished by receiving as input a tree corresponding to a source language text segment, and applying one or more decision rules to the received input to generate a tree corresponding to a target language text segment.

The present invention is a system of integrated, computer-based processes for monolingual information development. and multilingual translation. An interactive text editor enforces lexical and grammatical constraints on a natural language subset used by the authors to create their text, which they help disambiguate to ensure translatability. The resulting translatable source language text undergoes machine translation into any one of a set of target languages, without the translated text requiring any postediting.

A system and method for machine translation are disclosed. Source sentences are received. For each source sentence, a target sentence comprising target words is generated. A plurality of translation neighbors of the target sentence is generated. Phrase alignments are computed between the source sentence and the translation neighbor. Translation neighbors are scored with a translation scoring model, based on the phrase alignment. Translation neighbors are ranked, based on the scores. In training the model, parameters of the model are updated based on an external ranking of the ranked translation neighbors. The generating of translation neighbors, scoring, ranking, and, in the case of training, updating the parameters, are iterated with one of the translation neighbors as the target sentence. In the case of decoding, one of the translation neighbors is output as a translation. The system and method may be at least partially implemented with a computer processor.

The disclosed system translates into different languages HTML documents available through the World Wide Web. HTML documents are translated by machine translation software bundled in a browser. Alternatively, documents are retrieved as needed, translated, and stored on a Web server so user requests are serviced with a document that has been translated from a different language. The disclosed invention expands usage of the Internet for non-English speakers.

A machine translation method includes receiving source text in a first language and retrieving text fragments in a target language from a library of bi-fragments to generate a target hypothesis. Each bi-fragment includes a text fragment from the first language and a corresponding text fragment from the second language. Some of the bi-fragments are modeled as elastic bi-fragments where a gap between words is able to assume a variable size corresponding to a number of other words to occupy the gap. The target hypothesis is evaluated with a translation scoring function which scores the target hypothesis according to a plurality of feature functions, at least one of the feature functions comprising a gap size scoring feature which favors hypotheses with statistically more probable gap sizes over hypotheses with statically less probable gap sizes.

A translation method comprises: retrieving a fuzzy match text segment translation pair from a translation memory (TM) for an input source language text segment, the fuzzy match text segment translation pair comprising a fuzzy source language text segment having a fuzzy match to the input source language text segment and a corresponding translated target language text segment; extracting from the fuzzy match text segment translation pair an exact match phrase pair comprising a source language phrase that exactly matches a phrase of the input source language text segment and a corresponding translated target language phrase; and invoking a statistical machine translation (SMT) system to generate a proposed translation of the input source language text segment based on a statistical translation model that is enriched by the exact match phrase pair with the exact match phrase pair assigned a high statistical probability.

A system and method predict the translation quality of a translated input document. The method includes receiving an input document pair composed of a plurality of sentence pairs, each sentence pair including a source sentence in a source language and a machine translation of the source language sentence to a target language sentence. For each of the sentence pairs, a representation of the sentence pair is generated, based on a set of features extracted for the sentence pair. Using a generative model, a representation of the input document pair is generated, based on the sentence pair representations. A translation quality of the translated input document is computed, based on the representation of the input document pair.

The invention is a kind of Chinese and foreign languages double-way automatic translation method which is based on the sentential form conversion and adopts the Chinese phonetic codes with the word as the unit, and it belongs to the automatic translation technology field. By the method we can translate Chinese and another language such as English to each other expediently, on one hand, it overcomes the present faults that all the Chinese-foreign languages translation can only be done by the Chinese characters and Chinese spelling because of that Chinese uses the connecting written Chinese phonetic codes by word which only adopts twenty-six Latin letters as the code element and includes Chinese sound, rhythm and tone to expresses information when translating, that Chinese characters and Chinese phonetic signs can not be compatible with ASCII codes completely and the malpractice that before every translation, the Chinese information that Chinese characters expresses has to be divided to single word, on the other hand, because of adopting the sentential form conversion method which bases on the lexical and the syntax system coincident with the goal language, to make the automatic translation method more exact than the traditional translation method, and the effectiveness better. However the inputting Chinese form is Chinese characters, Chinese spelling or Chinese phonetic codes, it can not only translates by the traditional method but also translate after changing the Chinese characters to Chinese phonetic codes, likely, if needed, the Chinese phonetic codes gotten after translation can not only express the Chinese information directly, but also be conversed to Chinese characters, Chinese spelling, Chinese sound, Chinese special person, Chinese localism or minority language sound.

The disclosed system translates into different languages HTML documents available through the World Wide Web. HTML documents are translated by machine translation software bundled in a browser. Alternatively, documents are retrieved as needed, translated, and stored on a Web server so user requests are serviced with a document that has been translated from a different language. The disclosed invention expands usage of the Internet for non-English speakers.

A method and apparatus are provided for updating the vocabulary of a speech translation system for translating a first language into a second language including written and spoken words. The method includes adding a new word in the first language to a first recognition lexicon of the first language and associating a description with the new word, wherein the description contains pronunciation and word class information. The new word and description are then updated in a first machine translation module associated with the first language. The first machine translation module contains a first tagging module, a first translation model and a first language module, and is configured to translate the new word to a corresponding translated word in the second language. Optionally, the invention may be used for bidirectional or multi-directional translation.

Described herein are systems and methods for generating natural language sentences with Sequence-to-sequence (Seq2Seq) models with attention. The Seq2Seq models may be implemented in applications, such as machine translation, image captioning, and speech recognition. Performance has further been improved by leveraging unlabeled data, often in the form of a language models. Disclosed herein are Cold Fusion architecture embodiments that leverage a pre-trained language model during training. The Seq2Seq models with Cold Fusion embodiments are able to better utilize language information enjoying faster convergence, better generalization, and almost complete transfer to a new domain while using less labeled training data.

The invention relates to an audio-video recording and broadcasting device capable of translating speeches and marking subtitles automatically in real time for Chinese and foreign languages, and belongs to the technical field of voice and picture data processing equipment. The device comprises a microphone, a camera module 1, an audio-video synchronization signal marking module 2, a vocal language audio signal extraction module 3, a Chinese speech identification module 4, a machine translation module 5 for translating Chinese into foreign languages, an audio picture or an image picture subtitle stacking machine module 6, a foreign language speech synthesis module 7, an audio-video coding and compressing module 8, a network transmission module 9, a server module 10 provided with an audio-video decoding and decompressing software, a network transmission module 11, a client module 12 provided with audio-video play software, a foreign language speech identification module 13, a machine translation module 14 for translating the foreign languages into the Chinese and a Chinese speech synthesis module 15. The audio-video recording and broadcasting device can facilitate real-time communication of Chinese and foreign information.

The invention discloses a word and expression alignment method and device and relates to the technical field of machine translation. The word and expression alignment method and device achieve the technical goal of word and expression alignment. According to the technical scheme, the method comprises the steps of segmenting sentences, needing to be aligned, of two languages into words or expressions one by one, carrying out word and expression grouping, carrying out query matching in word and expression groups, and carrying out alignment of intertranslation pairs. The word and expression alignment method and device are used for accurate and complete building of a phrase list in the machine translation process.

The invention relates to a business logic code-free development method based on natural language understanding and conversion, and belongs to the technical field of code-free development. On one hand, development semantic extraction driven by natural language understanding is carried out; key development elements in the natural language description are automatically extracted to serve as a sequence labeling problem to be processed; and supervised training is performed on massive tagged code-free development demand corpora described by natural languages to obtain a development semantic extraction model. On the other hand, a natural language code generation task is is performed; after the intention of the user is analyzed, the metadata is sent into a model engine, and a corresponding code generation model is called by the model engine; and the code generation task is converted into a machine translation task. The method can reduce the software development and maintenance cost, improve the development efficiency, reduce the influence of demand change on the system and prolong the life cycle of the system.

The invention discloses a Japanese verb identification method for machine translation and belongs to the field of natural language processing. The method has the advantages that the method is based on rules and combines with dictionaries by analyzing the conjugated form rules of Japanese verbs, the verbs in a text can be identified completely, and basic forms of the verbs can be obtained by form restoration; the method can use common universal dictionaries and is high in adaptability and robustness; by the method, the lexical analysis accuracy and bilingual word corresponding effect in machine translation are increased, and the translation quality of machine translation is increased as a whole.

The invention provides a method and a device for aligning parallel spoken language material and a phonetic machine translation method and a system respectively adopting the alignment method and the device for parallel spoken language material. The alignment method of parallel spoken language material comprises the following steps: obtaining word alignment set based on a statistical method and a dictionary from the parallel spoken language material; conducting phrase alignment to the parallel spoken language the dictionary, so as to get phrase alignment set; and conducting word alignment in the alignment phrase in the parallel spoken language material, so as to get word alignment set based on the phrase alignment. The invention utilizes word alignment set with high accurate rate obtained from the parallel spoken language material in a corpus and based on the statistical method and the dictionary to conduct phrase alignment and further word alignment to the parallel spoken language material, so as to get phrase alignment set and word alignment set, as well as apply into use in phonetic machine translation, thereby reducing the ambiguity of spoken language alignment through utilizing the word completeness.

A machine translation method includes translating a source sentence using a first model, determining a back-translation probability of a translation result of the source sentence being back-translatedinto the source sentence using a second model, applying the back-translation probability to context information extracted from the source sentence in the first model, and retranslating the source sentence using the first model and the context information to which the back-translation probability is applied.

Computer systems, methods, and media for guaranteeing the quality of a language translation of content by using a computer network of human and / or machine translators and human reviewers of the original translation. The accuracy and quality of the original translation by a translator is ensured by a translation service provider computer system breaking the original translation into parts, and incorporating fabricated errors into a percentage of the parts while keeping the remaining parts unaltered. Incompetent reviewers are automatedly identified based on their ability to detect the system's fabricated errors. The system then performs an analysis on the remaining unaltered parts rated by the remaining competent reviewers to determine if the quality of the original translation meets a set standard. The translation service provider and the customer are not required to know: the target language; the accuracy of the human or machine translator; and / or, the accuracy of each translation reviewer.

A system and method for machine translation are disclosed. Source sentences are received. For each source sentence, a target sentence comprising target words is generated. A plurality of translation neighbors of the target sentence is generated. Phrase alignments are computed between the source sentence and the translation neighbor. Translation neighbors are scored with a translation scoring model, based on the phrase alignment. Translation neighbors are ranked, based on the scores. In training the model, parameters of the model are updated based on an external ranking of the ranked translation neighbors. The generating of translation neighbors, scoring, ranking, and, in the case of training, updating the parameters, are iterated with one of the translation neighbors as the target sentence. In the case of decoding, one of the translation neighbors is output as a translation. The system and method may be at least partially implemented with a computer processor.

The invention discloses a translation detection method and device and an electronic device, belonging to the technical field of language processing. The translation detection method comprises the following steps of obtaining a first translation, wherein the first translation is translated by an interpreter according to the original text; acquiring a plurality of second translations, the pluralityof second translations being translated for the original text by different machine translation engines; calculating a degree of matching between the first translation and the plurality of second translations; determining a target second translation according to the matching degree; determining the detection result of the first translation according to the target second translation. The invention determines whether the first translation has the translation result of the reference machine translation engine according to the target second translation, thereby effectively improving the reliabilityof the translation in the translation evaluation set, avoiding the occurrence of the unqualified translation in the translation evaluation set, facilitating the cost saving, and realizing the batch detection of the translation translated by the interpreter.

The invention discloses a machine translation model training method and device and a storage medium and belongs to the technical field of natural language processing. The method comprises the following steps: acquiring a multi-domain mixed training data set containing a plurality of training data pairs; performing data domain classification on the plurality of training data pairs to obtain at least two domain data subsets; determining at least two candidate optimization objectives of each domain data subset, and respectively training at least two candidate single domain models corresponding toeach domain data subset based on the at least two candidate optimization objectives; respectively testing at least two candidate single-domain models corresponding to each domain data subset, and selecting a candidate optimization target with the highest accuracy corresponding to the candidate single-domain model as a specified optimization target of the domain data subset; and training a hybriddomain model based on each domain data subset in the training data set and the corresponding specified optimization target. The quality of language data translation results of various fields by the hybrid field model can be improved.

The invention relates to a machine translation method for achieving the whole-sentence translation through preferential translation of idiomatic phrases composed of one or more phrase nests on the basis. According to the embodiment, the method comprises the following steps of marking idiomatic phrases, wherein idiomatic phrases in sentences in a source language are marked out; translating the idiomatic phrases, wherein the idiomatic phrases are divided into two parts to be translated respectively, and translations are regrouped; dividing short phrases, wherein remaining parts of original sentences are divided into all possible phrases, and the idiomatic parts are adopted as translated portions; constructing a candidate phrase list, wherein only phrases existing in a phrase translation probability list are screened out and added into the candidate phrase list; translating sentences, wherein the optimal translation is generated by means of an existing heuristic decoder and the candidate phrase list for the partially-translated source language sentences composed of the idiomatic translation and other non-translated parts. According to the first stage, the idiomatic phrases are translated, and according to the second stage, the remaining parts of the sentences are translated.

The invention discloses a training method of a machine translation model. The method comprises the steps of obtaining a first bilingual sentence pair; deleting punctuation marks in the first bilingualsentence pair according to a preset probability to obtain a second bilingual sentence pair; and training a machine translation model by using the second bilingual sentence pair. The technical effectsof reducing overlearning of a machine translation model on punctuation marks and improving translation accuracy are achieved. Meanwhile, the invention further discloses a training device of the machine translation model.

The invention provides a neural machine translation method based on similarity perception, and belongs to the technical field of natural language processing machine translation. The method comprises the following steps of firstly, creating a structure translation memory bank and a corresponding structure similarity algorithm; constructing a template translation memory bank and a corresponding template similarity algorithm; and then pre-identifying high-potential statements of character strings, structures and template dimensions in the test set; constructing multi-dimensional similarity prioriknowledge, and performing multi-dimensional similarity retrieval on all parallel statements in the training set. A posteriori regular objective function is utilized, discrete similarity priori knowledge is fused into a neural machine translation objective function, parameters of the priori knowledge are continuously iterated and updated, and a training process is guided. Finally, the multi-dimensional high-potential to-be-translated statements are translated by utilizing the trained neural translation model. The most similar sentences can be obtained in a finer-grained manner, and the reviewtime of human translators is shortened.

The invention provides a man-machine co-translation system based on exclusive engine turnover training, a man-machine co-translation method based on feedback training loop signals and a computer readable storage medium for realizing the method. According to the technical scheme, the system comprises a source text input component, and a semantic recognition module, an exclusive machine turning engine training ring, a man-machine translation window, a difference comparison interface and a reverse comparison translation engine which are connected with the source text input component. According tothe technical scheme, the result of machine translation can be utilized to the maximum extent, meanwhile, manual proofreading is conducted on translation results which may be wrong as much as possible, and the accuracy is improved while the efficiency is guaranteed, so that the accurate opportunity of manual translation intervention is provided, the translation efficiency can be guaranteed for large-scale corpus translation and translation occasions with high accuracy requirements, and meanwhile the translation accuracy is guaranteed.