從規(guī)則到生成：機(jī)器翻譯技術(shù)的演進(jìn)，現(xiàn)狀及未來(lái)發(fā)展趨勢(shì)

Abstract： With the rapid development of artificial inteligence and deep learning technologies， machine translation is playing an increasingly important role in facilitating cross-language communication.This paper systematically reviews the four evolutionary stages of machine translation technology， from the early rule-based systems， the statistical methods based on large-scale data， and the neural machine translation based on deep learning，to the current generative artificial intelligence （GenAl） translation models. It also shows that，although GenAl translation models have made significant progress in translation quality and efficiency，they still face problems such as data scarcity，limited model generalization ability， incomplete evaluation mechanisms，and lack of interpretability and ethical cultural sensitivity. This paper suggests that the future development of machine translation technology focus on enhancing the generalization abilityand interpretability of models， developing more comprehensive evaluation tools， and ensuring the cultural adaptability and ethical compliance of translation systems-all in seek of greater potential of machine translation ina wider range of application scenarios.

Key words： translation technology； machine translation ； generative artificial inteligence （GenAl）

1.引言

全球化進(jìn)程加速和數(shù)字化轉(zhuǎn)型深人使得不同語(yǔ)言和文化間的交流越來(lái)越頻繁，對(duì)高質(zhì)量機(jī)器翻譯技術(shù)的需求也愈發(fā)迫切（范夢(mèng)栩、皮姆，2021）。（剩余13483字）