METHOD AND DEVICE FOR VECTORIZING TRANSLATOR'S TRANSLATION PERSONALITY CHARACTERISTICS

摘要

The embodiments of the present disclosure provide a method and a device for vectorizing a translator's translation personality characteristics. The method comprises: constructing a bilingual sample set on the basis of historical translation corpus of a plurality of translators; inputting the bilingual sample set into a word vector model, and outputting word vectors; inputting each pair of bilingual corpus samples of the bilingual corpus sample set and the corresponding word vectors into an LSTM network-based encoder model and decoder model to perform training; and inputting the bilingual corpus set of each translator into the trained LSTM network to perform training, keeping the parameters of the encoder model unchanged, and obtaining the LSTM network corresponding to said translator; and generating a translator vector on the basis of the parameters of the decoder models in the LSTM network corresponding to said translator. According to the embodiments of the present disclosure, there is no need to filter the personality characteristics of a translator, and there is also no need to manually annotate sample data, so that the training cost is low and the accuracy is high, and the obtained translator vector can accurately and objectively reflect the translation personality characteristics of the translator.