You've probably already used an online machine translation service. You may turn to these translators with a certain frequency, even remembering how some translations, especially in the past, had a strange, often unnatural form to the native speaker and, not infrequently, unintelligible sentences. Despite obvious shortcomings, such resources have always been among the most popular on the Internet. Well, they are better than nothing, or rather than having to appeal to the time-consuming manual translation, when we simply needed to have at least a partial understanding of some foreign text.
Teaches us Wikipedia: "The idea of automatic translation dates back to the 17th century. In 1629, René Descartes proposed a universal language, in which equivalent ideas in different languages shared the same symbol. The Georgetown Experience (1954) included the completely automatic translation of more than 60 Russian phrases into English. The experience was a great success, and marked the beginning of an era in which research in automatic translation was abundantly funded. The authors considered that in three to five years automatic translation would be a trivial problem. "
Trivial problem? Decades and decades later, software translation is far from perfect. But the quality certainly no longer resembles that of the times of the Altavista translator. An online translation service has recently been launched by German-based company DeepL, which has impressed users. "Its translation tool is just as quick as the outsized competition, but more accurate and nuanced than any we’ve tried", said Devin Coldewey and Frederic Lardinois in an article on TechCrunch. DeepL's founder, Gereon Frahling, told them that they built a "neural translation network that incorporates most of the latest developments," to which the team added "their own ideas."
The Microsoft team, meanwhile, was quick to present its machine translation supported by artificial intelligence. They explain that "all of the machine translation products (websites or apps) available until late 2016 were based on algorithms using statistical methods to try to guess the best possible translation for a given word. This technology is called statistical machine translation." In the article "What is a neural network based translation?" they clarify one of its limitations: "it only translates words within the context of a few words before and after the translated word. For small sentences, it works pretty well. For longer ones, the translation quality can vary from very good to, in some cases, borderline nonsensical. It is almost always possible to see it has been machine-generated", they complete.
Microsoft has made available in the last days a page where you can enter a text and compare the versions of their statistical and neural translations.
You can test and compare these and various other tools that are in the match for the best automatic translations here. But it is likely that this race is still far from over, and it will take time to get to know his podium.
This text was fully translated from Portuguese text by a combination of automatic translation tools. ;)