55results about How to "The translation is accurate" patented technology

The invention discloses a machine translation method and device. The method comprises the steps that a sentence to be translated is received; at least one phrase fragment in the sentence to be translated is replaced by a preset character string, and a template which is matched with the sentence after replacing is searched from a template library after replacing operation every time; for the sentence matched with the template, a constant translation result of the part, corresponding to the constant of the template, of the matched sentence is acquired according to the template, a variable translation result of the part, corresponding to the variable of the template, of the matched sentence is acquired by decoding, and the constant translation result and the variable translation result are spliced; for the sentence which is not matched with the template in the template library, the translation result is acquired by decoding. According to the machine translation method and device, calculation can be reduced, and translation quality is improved.

The invention provides a word input method and an input method system having a translation function. The word input method comprises the following steps of: receiving a coded string input by a user; dissolving the coded string and obtaining a syllable unit sequence of a source language corresponding to the coded string; converting the syllable unit sequence of the source language and obtaining a syllable unit sequence of a target language; converting the syllable unit sequence of the target language and obtaining a transliteration result, and displaying the transliteration result; and outputting the appointed transliteration result according to an instruction of the user. In the word input method, a mapping relation between pronunciations of the two languages is constructed, so an effect that the user can obtain a correct spelling result through pronunciation can be achieved; under the help of the input method system, the user who forgets spelling can input a correctly spelled word or sentence; furthermore, some words which cannot be paraphrased or the paraphrase is not correct enough can be translated more accurately.

The invention discloses a continuous sign language video translation method based on the fusion of a time domain convolution network and a circulating neural network, which comprises the following steps: feature extraction of the sign language video and construction of a word list; processing of a time domain convolution network TCN; the processing of a bi-directional circulating neural network BGRU; word mapping process of feature; processing of a fusion network FL; optimization of network model parameters based on the fusion of the time domain convolution network TCN and the bi-directional circulating neural network BGRU; fusion and decoding of word coding vectors. The invention can effectively overcome the communication obstacle caused by inaccurate interpretation of sign language translation by utilizing the network structure of different perspectives, learn and analyze different manifestations of data by utilizing different networks, further improve the accuracy of sign language translation, and increase the robustness of sign language translation.

The embodiments of the invention provide a text translation model training method and device. The method comprises the steps that a first training text set is acquired, the first training text set comprises first-class corpus data and second-class corpus data, a first translation model is trained through the first-class corpus data, and after the first translation model meets the training stoppingcondition, a second translation model is trained through the second-class corpus data, wherein the first translation model and the second translation model share part of model parameters, and the number of second-class corpus data input during training of the second translation model is smaller than the number of first-class corpus data input during training of the first translation model; and, after the second translation model meets the training stopping condition, whether the first translation model continues to be trained or not is judged, and if it is determined that the first translation model does not continue to be trained any more, the trained first translation model serves as a text translation model. According to the embodiments of the invention, the translation result obtainedby training is more accurate and the translation style is clearer.

The invention discloses a zero parallel corpus multi-modal neural machine translation method, which comprises the following steps: pre-training a pre-constructed neural machine translation model by utilizing a source language with corresponding picture information and a target language monolingual corpus; translating sentences in the source language into sentences in a target language by using a pre-trained neural machine translation model; calculating a reward value of a sentence level according to an internal relation between the sentence obtained by translation and a corresponding picture, and updating pre-trained neural machine translation model parameters by adopting a strategy gradient reinforcement learning method by taking the maximum expected total reward value as an optimization target so as to obtain a trained neural machine translation model; and translating the given source language sentence by using the trained neural machine translation model. According to the zero parallel corpus multi-modal neural machine translation method, a neural machine translation model of a zero resource language pair can be established by utilizing the internal relation between pictures and characters.

The invention discloses a translation method and device. The translation method comprises the main steps that obtained voice signals are converted into audio information, then the audio information isseparated into multiple pieces of audio sub-information, and only the audio sub-information meeting the vocal print requirements is sent to a server for translation. By adopting the translation method, the translation work amount can be reduced, the translation efficiency can be improved, and the translation time can be shortened.

The embodiments of the invention provide a voice translation processing method and device, belonging to the technical field of language processing. The method includes the following steps: if that a first voice signal in the last round and a second voice signal to be translated in the current round correspond to the same language type is judged, adjusting the text data corresponding to the secondvoice signal based on the text data corresponding to the first voice signal to get the adjusted text data corresponding to the second voice signal; and acquiring a synthesized voice signal corresponding to the adjusted text data, and broadcasting the synthesized voice signal. As the first voice signal in the last round and the second voice signal to be translated in the current round may be correlated in content, the adjustment of the text data corresponding to the second voice signal based on the text data corresponding to the first voice signal can make the translation result better suit thescene and more accurate.

A cloud-based translation method for a mobile client is provided. The mobile client receives a translation instruction and sends translation request information to a cloud server based on the translation instruction, where the translation request information includes text information and translation type information. The mobile client receives a translation result returned from the cloud server. When the cloud server detects that the mobile terminal logins to an account, and a cloud glossary of the account contains a word entry corresponding to the text information, the cloud server returns the translation result in the cloud glossary. When the cloud server detects that the mobile terminal does not login to the account, or when the cloud server detects that the mobile terminal logins to the account but the cloud glossary does not contain the word entry corresponding to the text information, the cloud server translates the text information.

The invention provides an English lexical analysis method for neural network machine translation. The English lexical analysis method comprises the following steps: performing English word segmentation on a to-be-processed English text; carrying out query screening on each word obtained after English word segmentation by utilizing a special vocabulary; reserving result information of the queried words in a lexical analysis result; performing rule processing on the words which are not queried; carrying out word architecture restoration on the words meeting the rule processing conditions, and directly storing the words which do not meet the rule processing conditions in a lexical analysis result; and outputting a lexical analysis result, and adding the lexical analysis result into machine deep learning training. The invention further provides an English lexical analysis system for neural network machine translation. According to the English lexical analysis method for neural network machine translation, the problems that the machine learning efficiency is reduced and the translation quality is poor due to the fact that training corpus data of neural network machine translation is sparse can be solved.

The invention provides a classification model training method and device, a translation method and device and electronic equipment, and relates to the technical field of natural language processing and machine learning. The method comprises the following steps: acquiring bilingual parallel corpora, wherein the bilingual parallel corpora comprise a source language corpus and a target language corpus; determining translable fragments and untranslable fragments in the source language corpus, wherein the translable fragments are determined based on unambiguous fragments in the source language corpus; and training a classification model based on the translatable fragments and the untranslatable fragments. According to the above technical scheme, the classification model is trained based on the translable fragments and the untranslable fragments, the trained classification model is applied to translation, a translation result can be more accurate, and requirements for high quality and low time delay are met.

The invention discloses a Mongolian-Chinese machine translation method combining a Meta-KD framework and fine-grained compression. The method comprises the following steps: performing data preprocessing and data set division on Chinese corpora, English corpora and Mongolian corpora, learning Chinese-English translation by using the Meta-KD framework, training a BERT language model, learning a student model under the guidance of a meta-teacher according to a meta-distillation algorithm to obtain transferable knowledge for Mongolian-Chinese translation, and in combination with a fine-grained compression method, performing training verification of the Mongolian-Chinese translation on the student model. According to the method, data set training is performed through the Meta-KD framework, so that the method is more suitable for translation of small languages, and a more accurate translation result is obtained; and the fine-grained compression enables the trained model to have a higher training speed. Then, in combination with the fine-grained compression method, fine-grained compression is performed on information representation through information entropy, so that the purpose of model accelerated inference is achieved.

The invention discloses a scanning pen. The scanning pen comprises a pen body structure, a camera and a light beam positioning structure, the camera is used for collecting an image of a to-be-scannedinterface, the light beam positioning structure is used for emitting a light beam to form a positioning light mark on the to-be-scanned interface, and the positioning light mark is located on the leftside of a scanning window of the camera in the scanning direction of the scanning pen. The scanning pen is used; the positioning light mark can be formed through the light beam positioning structure;before scanning, the positioning light mark can be controlled to be located on the left side of a to-be-scanned character, which means that the scanning starting position is prompted to a user, the scanning window of the camera can be aligned with the to-be-scanned character, the complete information scanning can be realized based on the above basis, the conditions of missing the characters and the like are avoided, the finally scanned electronic information is more complete and accurate, and the translation result is more accurate.

The invention discloses a multifunctional English translator. The translator comprises a hand input unit, an infrared scanning unit, a speech input unit, a translation unit, a digital converter and a central processing unit; the hand input unit is used for inputting corresponding statements according to sentence pairs, terms or discourses and sending the input statements to the central processing unit; the infrared scanning unit is used for scanning English words, meanwhile converting the English words into scanning information and sending the scanning information obtained after conversion to the central processing unit; the speech input unit is used for inputting speech information through user accents, continuous speeches or an irregular word order and speed and sending the input speech information to the central processing unit and the digital converter; the digital converter is used for receiving the speech information sent from the speech input unit, converting the speech information into digital signals and sending the digital signals to the central processing unit. The translator is good in recognition effect, accurate in translation result, rich in function and quite convenient to use and intelligent.

The invention provides a translation method and system for a Chinese sentence containing unconventional vocabularies. For the Chinese sentence containing the unconventional vocabularies, firstly conventional vocabularies and the unconventional vocabularies in the Chinese sentence are identified through a word segmentation technology; and secondly two parts including the conventional vocabularies and the unconventional vocabularies are analyzed and translated, and a translation result is obtained through combination, wherein the conventional vocabularies can be obtained through an automatic translation tool, and for the unconventional vocabularies, a relatively accurate translation result can be obtained by adopting a corpus library retrieval technology proposed in the method. Through the method, a translator can obtain the translation result of the whole sentence only by finishing editing and proof-reading work with relatively small workload, and relatively high accuracy is ensured.

The invention provides a training method and device of a translation model, a translation method and device, equipment and a storage medium, and relates to the technical field of computers, in particular to the field of artificial intelligence such as natural language processing and machine translation. According to the specific implementation scheme, a sample document is processed, an RST chapter structure tree in a dependency form of the sample document is obtained, and the edge of the RST chapter structure tree in the dependency form represents the RST relation in the chapter of the sample document; determining an attention mechanism of the translation model to be trained based on an RST relationship in the RST text structure tree of the dependency form; and inputting the RST text structure tree in the dependency form and the sample document into the to-be-trained translation model for training to obtain a trained translation model. According to the embodiment of the invention, the translation result of the translation model is more accurate.