Application of GLR Algorithm in Statistical Machine Translation
Nan Liu · Procedia Computer Science · 2024
With the advancement of globalization and the development of multinational enterprises, more and more languages are being translated into each other. Statistical machine translation has the characteristics of high efficiency and low cost, and has broad application prospects. However, at present, the translation performance of Statistical machine translation system needs to be further improved. At present, a lot of work has proposed reordering methods to improve the translation performance of Statistical machine translation This paper proposes a reordering method of multiple candidate translations based on GLR algorithm, and uses GLR algorithm to improve the quality of Machine translation output translations. In order to verify the accuracy of GLR algorithm in reordering translations, experiments were carried out on the WMT19 translation automatic evaluation task. The experimental results show that this method can achieve the performance comparable to that of the automatic evaluation method for translations with reference translations, which shows that reordering in Statistical machine translation based on GLR algorithm can sort the quality of multiple output translations of the same source language sentence.