Neuron-based cache machine translation method and system

Abstract

The invention provides a neuron-based cache machine translation method. The method comprises the following steps of: submitting a translation request which comprises a translation text by a user; submitting the request to a background text processing server to carry out text preprocessing; carrying out punctuation and segmentation on the translation text in the translation request; searching whether an existing translation result exists in a cache or not by the background text processing server; carrying out corpus standardization on the translation text by the background text processing server; obtaining a preliminary translation result; generating a final translation result; and sending the final translation result to a front-end website server and storing the final translation result tothe cache. According to the method, a cache mechanism is added at a Service part of a neuron machine translation engine, when the translation request is received, search is carried out in the cache,and if history translation records exist in the cache, a result is directly output and the translation engine does not need to carry out re-translation, so that the resource utilization rate and translation efficiency of machine translation servers are greatly improved.

Description

technical field [0001] The invention relates to neuron machine translation technology. Background technique [0002] In the field of neuronal machine translation, the decoding time of the translation engine is much longer than that of traditional statistical machine translation. Since the decoding time directly affects the commercialization of machine translation systems, how to improve the translation efficiency of neuron machine translation engines is one of the most important research topics in the field of neuron machine translation. At present, the research on neuron machine translation engine is still in the category of decoder, but due to the training model and learning framework of neuron machine translation engine, even if the decoder is reconstructed, its decoding efficiency still has a lot of room for improvement. Contents of the invention [0003] The first technical problem to be solved by the present invention is to provide a neuron-based high-speed cache ma...

AI Technical Summary

Problems solved by technology

At present, the research on the neuron machine translation engine is still in the category of the decoder, but due to the training model and learning framework of the neuron machine translation engine, even if the decoder is reconstructed, its decoding efficiency still has a lot of room for improvement

Images