论文部分内容阅读
机器真的可以通过“学习”来取代人类的工作吗?实际上,优秀的译文应该具有灵活性和创造性,寄望于通过“大数据统计”路线来推进机器翻译的想法是不能的,因为这种翻译过程不包括对语句内容及其所属的言语行为的理解,这一重大缺陷也不能通过对该技术的改进而克服。例如,最近谷歌汉译英改用了神经网络技术,但是,尽管译文质量比以前的技术要好很多,可离人的翻译水平还差得很远。在此基础上,专家们对机器翻译的前景给出了
Is it really possible for the machine to replace human work through “learning”? In fact, good translations should be flexible and creative, with the expectation that the idea of advancing machine translation through the “Big Data Statistics” , Because this translation process does not include an understanding of the content of the sentence and the speech acts it belongs to. This major deficiency can not be overcome by improving the technique. For example, the recent Google Chinese to English conversion to neural network technology, however, although the quality of the translation is much better than previous technologies, the level of translation can be far behind. On this basis, experts have given the prospect of machine translation