In the myth about the tower of Babel the people decided to build a tower city that would get to heaven. And then the Creator realized that nothing would deter people and imagines they are not worth the bother. Then God created different languages to prevent people and that they could no longer easily work together. Nowadays, thanks to technology, we have unprecedented connectivity. However, we still live in the shadow of the tower of Babel. Language remains a barrier in business and marketing. Despite the fact that technological devices can easily connect people from different parts of the world often can’t.
Translation trying to catch: give presentations, contracts, instructions, outsourcing and advertising for everyone. Some agencies also offer so-called “localization”. For example, if a company goes to market in Quebec, it needs advertising on Quebec French, not European French. The company can suffer due to wrong translation.
Global markets are waiting for, but an English translation of the forces of artificial intelligence is not yet ready, despite the recent advances in natural language processing and sentiment analysis. The AI is still having difficulty with processing requests even in one language, let alone translation. In November 2016, Google has added a neural network in your translator. But some of her translations are still socially and grammatically strange. Why?
“To the credit of Google, the company has introduced quite a lot of improvements that have appeared almost overnight. But I don’t really use them. The language is difficult,” says Michael Housman, chief researcher, research RapportBoost.AI and lecturer, Singularity University.
He explains that the ideal scenario for machine learning and artificial intelligence will be fixed rules and clear criteria of success or failure. Chess is an obvious example, and together with them and th. Computer quickly mastered these games because the rules are clear and crisp, and the set of moves is limited.
“The language is almost the exact opposite. There is no clear and precise rules. The conversation can go in an infinite number of different directions. And of course you also need labeled data. You need to tell the machine what she’s doing is right and what is wrong.”
Hausman noted that the mark in the language of the information labels fundamentally difficult. “Two interpreters are unable to agree on the correct translation,” he says. “Language is Wild West from a data perspective”.
Google is now able to understand the proposals as a whole, without trying to translate individual words. But glitches still happen. Meifod Joerg, associate Professor of Spanish, specialist in Latin literature at the University of Jacksonville explains why accurate translations are not given artificial intelligence:
“The problem is that to understand the sentence as a whole is lacking. Just as the meaning of the word depends on the rest of the sentence (mostly in English), the meaning of a sentence depends on the rest of the paragraph and the text as a whole, and the value of the text depends on the culture, intentions of the speaker and other things. Sarcasm and irony, for example, have meaning only in a broader context. Idioms can also be problematic for computer-aided translation”.
“Google translation is a great tool if you use it as a tool that is not trying to replace human learning or understanding,” he says. “A few months ago I went to buy a drill at Home Depot and read the inscription under the car: “Saw machine”. (Machine saw). Below was the Spanish translation of ‘La máquina vió,’ which means “the Car is seen.” “Saw” is not translated as a noun and as a verb the past tense”.
Dr. Mahfud warns: “We need to know about the fragility of this interpretation. Because to translate is essentially to interpret, not just the idea but a feeling. Human feelings and ideas that can be understood only by people — and sometimes even we, the people, can not understand other people.”
He noted that culture, gender, and even age can create obstacles to that understanding, and the excessive dependence on technology leads to our cultural and political decline. Dr. Mahfud mentioned that the Argentine writer Julio Cortazar called dictionaries “cemeteries”. Automatic translators could be called a “zombie”.
Eric Cambria, academician, researching AI, and Professor of the Technological University of Nanyang in Singapore, is engaged in mostly natural language processing, which is the basis of translators on the basis of AI. As Dr. Mahfud, he sees the complexity and the associated risks in this direction. “There are so many things that we do unconsciously when we read the text”. Reading requires performing many unrelated tasks, which not under force automatic translators.
“The biggest problem of machine translation today is that we tend to pass from syntactic form of sentences in the language input to the syntactic form of this sentence in the target language. We humans don’t do that. We first decipher the meaning of a sentence in the input language, and then encode this value in the target language”.
In addition, there are cultural risks associated with these transfers. Dr. Ramesh Srinivasan, Director of the Laboratory of digital cultures at the University of California in Los Angeles, says that the new technological tools sometimes reflect underlying biases.
“There should be two parameters that determine how we design “intelligent systems”. One is values and, so to speak, biases of the one who creates the system. The second is a world in which the system will learn. If you are creating AI systems that reflect the biases of its Creator and into the wider world, sometimes spectacular failures”.
Dr. Sivanesan says that translation tools should be transparent in relation to opportunities and constraints. “You see, the idea that one system can take a language (which is very diverse semantically and syntactically) and combine them, or to what extent, to generalize, or even make one, this is ridiculous”.
Mary Cochran cofounder Launching Marketing Labs, see commercial growth potential. She noted that the lists on the online markets like Amazon can in theory automatically translate and optimize for customers in other countries.
“I believe that we now have only touched the tip of the iceberg, so to speak, about what AI can do with marketing. And with improved translation and globalization around the world the AI can not lead to the explosive growth of the market”.
Why is the AI still not mastered the translation languages?
Ilya Hel