Brief introduction of translation's prospect and obstruction

Do you want to communicate with foreigners without having to take a translator and just carry a portable digital machine with you? Even now it seems to be just wishful thinking. But maybe someday, we will be able to achieve a seamless communication with foreigners without having to have a headache for learning a foreign language.

Machine translation is a research area with many challenges, which is because it is difficult to achieve, and it has been ranked as the top ten scientific and technological problems in the 21st century. Development so far, translation has gone through countless difficult and tortuous road of development, which not only have success and excitement, but also full of frustration and confusion.

However, demand is always ahead of the application. With the acceleration of globalization and the increasing frequency of international communication in recent years, there is an unprecedented increase in the demand for translation. And the universal application of the Internet makes online translation the play of today's translation.

In this field, competition is becoming more and more intense. Today's powerful, Easy-to-use online translator tools have Google translation, Bing translation, Facebook translation, baby fish translation, Babylon translation, among them, the star of Google translation of the most unique, at the same time the most representative.

Google translation currently offers real-time translations between 63 major languages, and it can provide translations between any two languages that are supported, including words, sentences, text, and Web pages. It also helps users read search results, Web pages, emails, YouTube video subtitles, and other information, and users can even have real-time multilingual conversations within Gmail.

Google translation is mainly the use of statistical translation model, to the calculator to enter a large number of text text, including the source language text, as well as the corresponding target language manual translation of text, through massive statistical data to improve translation accuracy.

One important reason for using a statistical translation model is that Google's translation employs a cloud computing architecture. The architecture has a distributed computing system (MAPREDUCE) and distributed Storage Systems (BigTable) developed by Google, which are both creative and hugely scalable, making Google competitive in system throughput.

Translation's more exciting application lies in the real-time translation of everyday conversations. This area is also Google's lead; it has strong speech recognition technology, can be automatically retrieved through the sound, and then the speech recognition and translation together.

In November 2011, Google launched a new mobile phone translation software, which supports 14 languages including Mandarin Chinese. Users who speak to Google's Android smartphone can almost instantly hear their source language being translated into the target language, while the language of the caller is translated into the user's mother tongue.

Google chairman Schmidt said the mobile translation service will be greatly improved, and may even be "simultaneous interpretation", the outlook is optimistic.

Google's chief scientist, Asch Weinugoppal, says Google's ultimate goal is to develop mobile software that translates into 300 languages spoken by at least 1 billion people worldwide.

In addition to the traditional translation ideas, a new method known as the decryption of foreign languages also exposes the sharp corners.

The new translation method, proposed by scientists at the University of Southern California, treats English as an initial language, and a foreign language that needs to be translated resembles an encrypted, advanced text that decodes and translates a foreign language into English.

But the reality is bone. At present, this translation method can only translate some short sentences or film subtitles, but it has not been applied to long sentence or fragment translation, but it has opened a new way of thinking for translation research and can be applied to any language. The scientists say the translation method can even decipher "dolphin tone" or "alien language" because it does not need any existing text.

In fact, either way, the biggest factor affecting translation development is the quality of the translation. The quality of translation is far from the ultimate goal in terms of what has been achieved.

Zhouhaizhong, a Chinese mathematician and linguist, has pointed out in the paper, "50 Years of Machine translation": To improve the quality of translation, the first thing to be solved is the language itself rather than the programming problem, and it is certainly impossible to improve the translation quality by relying on several programs to do the translation system.

At the same time, he also pointed out: "The human brain is how to carry out the fuzzy recognition of language and logical judgment," translation to achieve the "letter, Tatsu, ya" degree is impossible. This is also a bottleneck restricting translation quality improvement.

It is worth mentioning that the American inventor, futurist Rekozwil, recently interviewed by The Huffington Post, predicted that by 2029 the quality of translation would reach the level of human translation. There are many controversies in the academic circles about this argument.

No matter what, at present is people to translation the most optimistic period, this kind of attention is based on an objective understanding and rational thinking. We also have reason to believe that, with the joint efforts of calculator experts, linguists, psychologists, logics and mathematicians, the bottleneck of translation will be solved.