FIELD: machine translation.
SUBSTANCE: invention can be used to perform context-sensitive translation. The method comprises generating an augmented sequence of input tokens based on an input sentence in the first language and a context word in the first language inserted into the input sentence. The context word is represented as an input token in an augmented sequence of input tokens, placed at a predetermined position and containing contextual information. The method further comprises iteratively generating a sequence of output tokens based on the augmented sequence of input tokens, wherein the sequence of output tokens comprises an output token placed at a predetermined position and representing a corresponding context word in the target language, and another output token representing a context-sensitive translation of the word from input sequence.
EFFECT: increased translation accuracy.
28 cl, 7 dwg
| Title | Year | Author | Number |
|---|---|---|---|
| METHOD AND SERVER FOR PERFORMING PROBLEM-ORIENTED TRANSLATION | 2021 |
|
RU2820953C2 |
| METHOD AND SYSTEM FOR RECOGNIZING USER'S SPEECH FRAGMENT | 2021 |
|
RU2808582C2 |
| METHOD AND SERVER FOR TRAINING MACHINE LEARNING ALGORITHM FOR TRANSLATION | 2020 |
|
RU2789796C2 |
| METHOD AND SYSTEM FOR TRAINING CHATBOT SYSTEM | 2023 |
|
RU2820264C1 |
| METHOD AND SERVER FOR PROCESSING TEXT SEQUENCE IN MACHINE PROCESSING TASK | 2020 |
|
RU2775820C2 |
| METHOD AND SERVER FOR TRAINING MACHINE LEARNING ALGORITHM IN TRANSLATION | 2020 |
|
RU2770569C2 |
| METHODS AND ELECTRONIC DEVICES FOR PACKAGING REQUESTS INTENDED FOR PROCESSING BY PROCESSING UNIT | 2021 |
|
RU2810916C2 |
| METHOD OF TRAINED RECURRENT NEURAL NETWORK DEBUGGING | 2019 |
|
RU2715024C1 |
| MULTISTAGE TRAINING OF MACHINE LEARNING MODELS FOR RANKING SEARCH RESULTS | 2021 |
|
RU2824338C2 |
| SYSTEM AND METHODOLOGY OF AUTOMATIC LANGUAGE LEARNING ON BASIS OF SYNTACTIC MODELS FREQUENCY | 2015 |
|
RU2632656C2 |
