Article Preview
Top1. Introduction
In recent years, neural machine learning that uses artificial neural network has witnessed significant development and grown in such a refractory manner that is indescribable with words. Neural machine learning that has been designed in similar patterns to the neurons of a human brain, has increased fluency & accuracy in language translation, and enabled the system itself to learn from millions of examples, in aims to provide the best solutions to many problems in image recognition, speech recognition as well as natural language processing.
Neural machine translation has been introduced as a promising approach to address the many deficiencies of machine translation systems. The public debut of Google Translate in 2006 utilized machine translation systems. In September of 2016, Google Translate fully switched to neural machine translation with the potential of addressing many shortcomings of machine translation (Sutskever, Vinyals & Le, 2014).
By using neural machine learning or neural Artificial Intelligence, and incorporating techniques such as rewriting-based paradigms and on-device processing, neural machine translation has taken quantifiable leaps in language translation accuracy (Wu et al., 2016; Chan, 2020; Wiggers, 2020). In 2016, Google Neural Machine Translation had only less than ten languages available, but this has increased to over 100 languages by 2020 (Statt, 2020).
Neural machine learning is a branch of AI that concerns the use of software to translate text from one language to another. Neural machine learning is also an end-to-end learning approach for automated translation, with the potential to overcome many of the drawbacks of conventional phrase-based translation systems (Wu et al., 2016). Although neural machine learning is still limited in some capacities, its rapid improvements and developments are contributing to an explosion of digital content, which in turn, is contributing to the rapid development of globalization including international trade.
When Google first launched its language translation service called Google Translate on April 28 of 2006 using Machine Learning (ML), it was only able to support two and hundreds of users (Turovky, 2016; Makadia, 2018). By November 2016, after Google’s development of the neural machine translation system called the Google Neural Machine Translation, it is now able to support over 100 languages and has hundreds of millions of users. Google Neural Machine Translation is an artificial neural network which uses deep machine learning to mimic the functioning of a human brain and increases fluency as well as accuracy (Turovsky, 2016; Schuster et al., 2016; Wu et al., 2016; Sen, 2016; Turner, 2016). As of 2019, fourteen other languages, although are not yet supported by Google Translate, are available in the Translate Community (Translate Community, 2019).
Anecdotal evidence points towards other multilingual cyberspace in the workings including the surge of domestic search engines are dominating some countries and Google, which dominates up to 92.06% of the world market share. Such insurgence of neural machine learning technologies in the World Wide Web allows users from all tiers to “cross cyber borders” without having to comprehend a local language of that country (Philips, 2018; StatsCounter, 2020). Figure 1 shows the global market share of the world’s top search engines in 2019. Microsoft’s Bing take up 2.41% of the world market share, with Yahoo! coming in third place at 2.07%, and other search engines such as Baidu, Yandex, Ask, DuckDuckGo and AOL, dominating less than 1.01% of the world market share.
Figure 1.
Global market share of world’s top search engines