Google Translate is a free, multi language translation service developed by Google, to translate text. It offers a website interface, mobile apps for Android and iOS, and APIs that help developers build browser extensions and software applications. Google Translate supports more than 100 languages ​​at various levels and starts in May 2017, serving more than 500 million people every day.
Launched in April 2006 as a statistical machine translation service, it uses the United Nations and transcripts the European Parliament to collect linguistic data. Instead of translating the language directly, first translate the text into English and then into the target language. During translation, he looks for patterns in millions of documents to help break the best translations. Accuracy has been criticized and ridiculed several times. In November 2016, Google announced that Google Translate would switch to a neural machine translation engine - Google Neural Machine Translation (GNMT) - which translates "whole sentences at once, not just piece by piece.This uses this broader context to help find out the most relevant translation, which is then rearranged and adjusted to make it more like talking a human with the proper grammar. " Initially only enabled for multiple languages ​​by 2016, GNMT is gradually being used for more languages.
Video Google Translate
Features
Google Translate can translate various forms of text and media, including text, greeting, images, sites, or real-time video, from one language to another. It supports more than 100 languages ​​at various levels and in May 2017, serving more than 500 million people every day. For some languages, Google Translate can pronounce translated text, highlight appropriate words and phrases in source and target text, and act as a simple dictionary for one-word input. If "Language detection" is selected, text in an unknown language can be automatically identified. If a user enters a URL in the source text, Google Translate will generate a hyperlink to the machine translation of the website. Users can save translations in "phrase books" for later use. For some languages, text can be entered via the on-screen keyboard, via handwriting recognition, or speech recognition.
Browser integration
Google Translate is available in some web browsers as a downloadable optional extension that can run a translation engine. In February 2010, Google Translate was integrated into the Google Chrome browser by default, for automatic optional webpage translation.
Mobile app
The Google Translate app for Android and iOS supports more than 100 languages ​​and can translate 37 languages ​​via photo, 32 by voice in "conversation mode", and 27 via real-time video in "augmented reality mode".
The Android app was released in January 2010, and for iOS on February 8, 2011.
Android version January 2011 experimented with "Conversation Mode" which allows users to communicate seamlessly with nearby people in other languages. Originally limited to English and Spanish, this feature received support for 12 new languages, still in testing, the following October.
In January 2015, the app gained the ability to translate physical signs in real time using device cameras, as a result of Google's acquisition of the Lens Firm application. The original January launch only supported seven languages, but the July update added support for 20 new languages, and also improved the speed of the Conversation Language translation.
API
In May 2011, Google announced that Google Translate API for software developers has been discontinued and will stop functioning. The Translation API page states the reason as "substantial economic burden caused by extensive abuse" with an end date set for December 1, 2011. In response to public pressure, Google announced in June 2011 that the API will continue to be available as a service charge.
Because APIs are used on many third-party websites and apps, the original decision to write them off caused some developers to criticize Google and question the viability of using Google APIs in their products.
Google Assistant
Google Translate also provides translations for Google Assistants and devices run by Google Assistants such as Google Home and Google Pixel Buds.
Maps Google Translate
Supported languages ​​
The following languages ​​are supported in Google Translate.
Translation methods
In April 2006, Google Translate was launched with a statistical machine translation engine.
Google Translate does not apply grammatical rules, because the algorithm is based on statistical analysis rather than traditional rule-based analysis. The original creator of the system, Franz Josef Och, has criticized the effectiveness of rule-based algorithms that support the statistical approach. It is based on a method called statistical machine translation, and more specifically, on research by Och who won the DARPA contest for speed machine translation in 2003. Och is the head of the Google machine translation group up out to join Human Longevity, Inc. in July 2014.
According to Och, a solid basis for developing a statistical machine translation system that can be used for a pair of new languages ​​from scratch will consist of bilingual text corporations (or parallel collections) of more than 150-200 million words, and two monolingual corporates each more than one billion words. The statistical model of this data is then used to translate between the languages.
To obtain this vast amount of linguistic data, Google uses the transcript of Parliament of the United Nations and Europe.
Google Translate does not translate from one language to another (L1 -> L2). Instead, it is often translated first into English and then into the target language (L1 -> EN -> L2).
When Google Translate generates translations, it searches patterns in hundreds of millions of documents to help break the best translations. By detecting patterns in documents translated by human translators, Google Translate makes a smart guess what translation should be.
Prior to October 2007, for languages ​​other than Arabic, Chinese and Russian, Google Translate was based on SYSTRAN, a software engine still used by some other online translation services such as Babel Fish (now inactive). Since October 2007, Google Translate has used its own internal technology based on statistical machine translation instead.
Google Neural Machine Translation
In September 2016, a research team at Google led by software designer Harold Gilchrist announced the development of a Google Neural Machine Translation (GNMT) system to improve fluency and accuracy in Google Translate and in November announced that Google Translate would switch to GNMT. (wikipedia)
The Google Translate nerve machine translation system uses a large end-to-end artificial neural network capable of deeply learning, in particular, long-term, short-term memory networks. GNMT improves the quality of the translation because it uses an example-based machine translation method (EBMT) where the system "learns from millions of examples." It translates "the whole sentence at a time, not just piece by piece." He uses this broader context to help find the most relevant translations, which are then rearranged and adjusted to be more like talking human beings with proper grammar. " GNMT's "proposal architecture" of "system learning" was first tested in over a hundred languages ​​supported by Google Translate. With a tip-to-end framework, "the system learns from time to time to create better and more natural translations." The GNMT network is able to translate interlingual machine translations, which encode "semantic sentences rather than simply memorize phrase-to-phrase translation," and the system does not create its own universal language, but uses "commonality found among many languages". GNMT was first activated for eight languages: to and from English and Chinese, French, German, Japanese, Korean, Portuguese, Spanish, and Turkish. In March 2017, the language was activated for Hindi, Russian, and Vietnamese, followed by Indonesian, Bengali, Gujarati, Kannada, Malayalam, Marathi, Punjabi, Tamil, and Telugu in April.
GNMT has the ability to translate directly from one language to another (L1 -> L2), which improves on the previous version of Google Translate which was first translated into English and then into the target language (L1 -> EN -> L2). The GNMT system is also capable of translating Zero-Shot - translating between language pairs (eg, Japanese to Korean) that "the system was never explicitly seen before."
Crowdsourcing
In 2014, Google launched the "Community Translation", a platform aimed at improving translation services by seeking help from volunteers. In August 2016, the Google Crowdsource app was released to Android users, where a translation assignment was offered.
Limitations
Some languages ​​produce better results than others. Google Translate is a good performer especially when English is the target language and its source language is from the European Union because of the prominence of translated EU parliamentary records. The 2010 analysis shows that French to English translations are relatively accurate. However, if the source text is shorter, rule-based machine translation often performs better; this effect is very clear in the Chinese to English translation. While editing translations may be submitted, in Chinese specifically one can not edit the sentence in its entirety. Instead, someone has to edit an arbitrary character set sometimes, leading to incorrect editing.
Reviews
Shortly after launching the translation service for the first time, Google won international competition for English-Arabic and English-Chinese translation.
Errors and strangeness of translation
Because Google Translate uses statistical matching to translate, translated text can often include seemingly unreasonable and obvious errors, sometimes swapping common terms for common similar terms but no similarities in other languages, or reversing the meaning of a sentence. New websites like Bad Translators and Translating Parties have used this service to generate funny texts by translating back and forth between multiple languages, similar to whispering Chinese children's games.
Court use
In 2017, Google Translate was used during court hearings when a court official at the Teesside Magistrates Court failed to order an interpreter for a Chinese defendant.
See also
References
External links
- Official website
- The Translation Community
Source of the article : Wikipedia
0 Comments