Non-autoregressive neural machine translation method and device, computer device and medium

Abstract

The embodiment of the invention discloses a non-autoregressive neural machine translation method and device, a computer device and a medium. The method comprises the steps of obtaining a source sentence of a source language and a word vector corresponding to a word in the source sentence; encoding the word vector corresponding to the word to obtain an encoding vector of the concerned context information; determining a to-be-translated sentence according to the source sentence, wherein the to-be-translated sentence comprises a to-be-translated word; according to the word vectors corresponding to the words to be translated and the encoding vectors, reordering the words to be translated in the sentence to be translated according to the structure of a target language to obtain a pseudo-translated sentence; translating the pseudo-translated sentence into a target sentence of the target language according to a word vector corresponding to a to-be-translated word in the pseudo-translated sentence and the encoding vector; outputting the target sentence. The scheme can improve the translation quality.

Description

Technical field [0001] This application relates to the field of artificial intelligence technology, in particular to a non-autoregressive neural machine translation method, device, computer equipment and medium. Background technique [0002] Artificial Intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology of computer science, which attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can react in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines, so that the machines have the functions of perception, reasoning and decision-making. ...

AI Technical Summary

Problems solved by technology

[0004] However, there is still a certain gap between the translation quality based on NAT technology translation and autoregressive NMT (Neural Machine Translation, neural machine translation) technology. At present, most NAT technologies have the problem of large search space when decoding, that is, when the NAT model predicts the current word The surrounding words need to be inferred to determine which part of the source sentence the current word should translate, and also which correct word that part should be translated into

Therefore, the translation quality is lower

Images