An Analysis of Post-editing Efficiency Based on Online Translation Tools in English-Chinese Translation毕业论文

 2021-04-02 11:04

摘 要

随着全球化进程加快,世界范围内翻译需求激增,语言服务行业发展迅速,一些新的研究领域也因此得到了学界的重视。其中,得益于计算机技术的发展,机器翻译这个新兴领域受到了广泛关注。经历了基于规则的机器翻译和统计机器翻译后,神经网络翻译这项涉及计算机科学、语言学、数学、神经认知科学等多个学科的技术引领机器翻译进入了新时代,大幅提升了翻译质量。今年4月,博鳌论坛首次采用腾讯AI同传、GMIC大会采用搜狗同传更是将神经网络翻译推向高潮。但同时,这项技术也给翻译市场带来了巨大冲击,引发了业内担忧:机器翻译是否会取代人工翻译?本文将着重介绍基于规则的机器翻译、统计机器翻译和神经网络机器翻译的发展,分析其优劣势及其对人工翻译的影响,探讨机器翻译未来要解决的问题和译后编辑的必要性,以期对机器翻译和译后编辑紧密结合形成新的模式提供一些参考。

关键词:神经网络翻译;译后编辑;人工翻译;统计机器翻译;基于规则的机器翻译

Contents

1 Introduction………………………………………………………..………….……....…………7

2 Demonstration…………………………………………………………………….....…………10

2.1 History of Machine Translation ……………………………….……………………….....…10

2.1.1 Rule-based Machine Translation………………………………….……………….………10

2.1.2 Statistical Machine Translation…………………………………………………….....……11

2.1.3 Neural Machine Translation…………………………………………………….….………11

2.2 Mechanism of Neural Machine Translation…………………..…………………………...…11

2.3 Advantages of Neural Machine Translation…………………..…………………………...…12

2.3.1 Convenience and Affordability ……………………….…………………………………...12

2.3.2 Accuracy and Mass Storage …………………………..………………………………...…13

2.4. Disadvantages of Neural Machine Translation…………………………………………...…13

2.4.1 Language Mismatches……………………...…………………………………………...…14

2.4.2 Out-of-vocabulary Word Ambiguity…………..………………………………………...…15

2.4.3 Difficulty in Model Training and Feature Addition……………………………………..…15

2.4.4 Lack in Translation Diversity………………………………………………………...……16

2.5 Impact on Human Translation………………………………………………………......……16

2.5.1 Overview of Post-editing……………………………………………………………..……17

2.5.2 Necessity of Post-editing………………………………………………………………..…17

2.5.2.1 Personalized Demand amp; Privacy Protection………………..………….............……...…17

2.5.2.2 Ethic Codes…………………………….………………………………………….…..…18

2.5.2.3 Culture amp; Civilization………………………………………………………………...…18

3 Conclusion…………………………………………………………………………………..…18

Works Cited…………………………………………………………………………...........……21

Bibliography……………………………………………………………………………..………22

  1. Introduction

The past few decades have witnessed not only rapid development of technology and globalization but also the exceedingly increasing demand for translation worldwide and the booming of language service industry. What is known to us all is that long time and high price have, to some extent, prohibited the widespread use of human translation, leaving it unable to fully meet the needs of such a gigantic translation market. The emergence of machine translation that by automatically translating texts from one natural language to another through computer increases translation speed while reducing costs, conforms to the needs of the times and has been welcomed by people.

After statistical machine translation, Baidu introduced its brand new neural network machine translation system in 2015, sparkling a flurry in the industry to build up comparable ones. In 2016, Google released its commercial neural translation system with translation accuracy increased by 58%- 87% compared with its previous phrase-based machine translation system (Wu et al.), which actually promoted growth in the usage and popularity of neural machine translation. Last year, NetEase also launched its neural translation system and portable translator, followed by iFLY TEK, Huawei and Microsoft.

In April 2018, machine translation became a public hotspot once again as Tencent AI simultaneous interpretation service, covering real-time interpretation screen-casting at conference site, WeChat applet live broadcasting and conference minutes playback, publicly supported the sub forums at Boao Forum for Asia. (However, as Boao Forum is a high-level forum modeled after the World Economic Forum, the AI translation service was thus subject to human interpreters’ review in the opening ceremony.) According to Tencent’s director in charge Li Xuezhao, Mr. Translator was capable of translating 15 languages; its average speech recognition accuracy at conferences had reached 97% and the acceptability was over 93%. Although Tencent had provided bilingual interpretation solutions for large-scale conferences such as its Media Summit, Global Partner Conference and Regional Influence Event, it was still a challenge to perfectly accomplish high-quality interpretation in such a special scene with high level of professionalism and wide coverage in topics at Boao Forum. Different from previous single-speaker conferences, the forum gathered over 2,000 senior government officials, business executives and leading academics of various professional backgrounds who discussed in a large variety of languages (random in Chinese and English). When business elites from all over the world delivered brilliant speeches on economic development in the future, on both sides of the venue cast the recognized speeches and translated subtitles in real time. At the same time, the viewers listened to the bilingual interpretation of the guests’ speeches and could review the records via WeChat applet.

您需要先支付 80元 才能查看全部内容!立即支付

课题毕业论文、开题报告、任务书、外文翻译、程序设计、图纸设计等资料可联系客服协助查找,优先添加企业微信。