A method for translating a text written or otherwise communicated in a source natural language into a text written or otherwise communicable in target natural language, in reliance upon a multidimensional model, relies on determining the core concept in the sentences of the source text, and leverages the determined core concepts to create the target language translation. The method includes processing the source natural language text into sentences, then parsing the sentences, including assigning codes and/or directional operators to realize parsed sentences according to the model. The sentence models are then processed effect the actual translation to the target natural language text, and communicated.

Various aspects of the subject technology relate to systems, methods, and machine-readable media for language translation based on image search similarities. These aspects include an image retrieval system using a convolutional neural network that is trained to identify a correlation between an image and a language term, and using an image search engine to search against images corresponding to visual words that are responsive to a given search query in a given spoken language. These aspects include access to interaction probability data that identifies user interaction probabilities for the visual words to determine a correlation between the input language terms of the search query and the rate at which users interact with images of a corresponding visual word that is responsive to the search query. The system then provides a prioritized listing of images that is responsive to the given search query based on the identified user interaction probabilities.

Systems and methods are provided for Cross-lingual Transfer Interpretation (CTI). The method includes receiving text corpus data including premise-hypothesis pairs with a relationship label in a source language, and conducting a source to target language translation. The method further includes performing a feature importance extraction, where an integrated gradient is applied to assign an importance score to each input feature, and performing a cross-lingual feature alignment, where tokens in the source language are aligned with tokens in the target language for both the premise and the hypothesis based on semantic similarity. The method further includes performing a qualitative analysis, where the importance score of each token can be compared between the source language and the target language according to a feature alignment result.

A technique for translating text strings includes receiving a source language text string from an application, determining that a translated text string that includes a translation in a target language of the source language text string is not available for use by the application, transmitting the source language text string to a translation service for translation, receiving the translated text string from the translation service, and causing the translated text string to be available for use by the application.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for adapting a language model are disclosed. In one aspect, a method includes the actions of receiving transcriptions of utterances that were received by computing devices operating in a domain and that are in a source language. The actions further include generating translated transcriptions of the transcriptions of the utterances in a target language. The actions further include receiving a language model for the target language. The actions further include biasing the language model for the target language by increasing the likelihood of the language model selecting terms included in the translated transcriptions. The actions further include generating a transcription of an utterance in the target language using the biased language model and while operating in the domain.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for adapting a language model are disclosed. In one aspect, a method includes the actions of receiving transcriptions of utterances that were received by computing devices operating in a domain and that are in a source language. The actions further include generating translated transcriptions of the transcriptions of the utterances in a target language. The actions further include receiving a language model for the target language. The actions further include biasing the language model for the target language by increasing the likelihood of the language model selecting terms included in the translated transcriptions. The actions further include generating a transcription of an utterance in the target language using the biased language model and while operating in the domain.

A machine translation method includes: receiving a sentence, the sentence including a plurality of words; calling a machine translation model obtained through training, the machine translation model including a partitioning model and a translation model; partitioning the sentence based on the partitioning model and according to word vectors of the words, to obtain to-be-translated blocks, each to-be-translated block including at least one of the words; and translating the sentence based on the translation model and the to-be-translated blocks, to obtain a translation result.

Methods and electronic devices for executing offline translation of a source word into a target word via a Neural Network an encoder and a decoder. The method includes splitting the source word into input tokens, generating vector representations for input tokens, and generating a first sequence of output tokens representative of a first candidate word. In response to the first candidate word not respecting at least one pre-determined rule, the method includes triggering the decoder to generate a second sequence of output tokens having a different at least one last output token than at least one last output token of the first sequence. The second sequence is representative of a second candidate word. In response to the second candidate word respecting the at least one pre-determined rule, the method includes determining that the second candidate word is the target word.

A data processing system for generating training data for a multilingual NLP model implements obtaining a corpus including first and second content items, where the first content items are English-language textual content, and the second content items are translations of the first content items in one or more non-English target languages; selecting a first content item from the plurality of first content items; generating a plurality of candidate labels for the first content item by analyzing the first content item with a plurality of first English-language NLP models; selecting a first label from the plurality of candidate labels; generating first training data by associating the first label with the first content item; generating second training data by associating the first label with a second content item of the second content items; and training a pretrained multilingual NLP model with the first training data and the second training data.

Methods and servers for training a translation model for translation between a rare language from a group and a target language. The method includes acquiring an actual example of translation and using a transliteration function for generating a synthetic actual example of translation. The method includes acquiring a sentence in the target language, generating an artificial translation of that sentence using back-translation, and thereby generating a given artificial example of translation. The method includes generating a synthetic artificial example based on the given artificial example. The method includes training the translation model based on the synthetic actual example of translation and the synthetic artificial example of translation.