Neural networks are machines and algorithms developed to behave like the human brain--but a development from Google Translate shows that (once again) AI can outperform humans in a big way. Google's AI can now translate language pairs it has not been trained for. To be clear, this means that it can translate between languages that it wasn't taught to translate. This works if the AI first translates both of the languages into a common language that it knows. The development is detailed in a paper published on Cornell University's arXiv.

Google says its artificial intelligence has taught itself to'translate between languages that it doesn't even know' 'Zero-shot translation' can translate between languages it doesn't know Deep-learning researchers developed Google Neural Machine Translation GNMT developed algorithm that'self-teaches' it to translate languages'Zero-shot translation' can translate between languages it doesn't know GNMT developed algorithm that'self-teaches' it to translate languages Google headquarters in Menlo Park, California is seen in the above stock photo. Google says it has built an algorithm that allows its Google Translate service to translate languages it doesn't even know Google says that its artificial intelligence uses a'token' at the beginning of the input sentence to specify the required target language to translate to Skating on thin ice: Wife of Vladimir Putin's spokesman... 'This had nothing to do with Donald': Rosie O'Donnell... Liberty University President Jerry Falwell Jr. says he... New Orleans violence'out of control' says mayor after one... Skating on thin ice: Wife of Vladimir Putin's spokesman... 'This had nothing to do with Donald': Rosie O'Donnell... Liberty University President Jerry Falwell Jr. says he... New Orleans violence'out of control' says mayor after one... Man punches kangaroo in the face to save dog being strangled CCTV shows woman hit by a CAR as thieves dodge paying for petrol Mob storm police station and lynch suspected paedophile Moment judge gets slapped in the face at 2016 IFBB Diamond Cup Watch the original 1972 trailer for'Last Tango in Paris' Heart-stopping moment child climbs on tracks in front of train Christmas advert so touching it's being shared around the world Adorable baby'sings' along with Dutch champion beatboxer B-Art Which body types do women like best? Christmas advert so touching it's being shared around the world Adorable baby'sings' along with Dutch champion beatboxer B-Art Which body types do women like best? 'Don't believe a thing you hear, unless it comes from me':... EXCLUSIVE: Man who punched out kangaroo is a ZOOKEEPER:... EXCLUSIVE: The'arrogant' man behind illegal artist enclave... EXCLUSIVE: Ex-American Idol contestant Corey Clark describes... Pictured: Twelve victims identified as firefighters recover... Is Europe's Brexit revolution over? Gloating left-wing... High school algebra teacher calls off her wedding after she... 'Roo want a piece of me?

A new Google Translate system of artificial intelligence called Zero Shot Translation can now create translations between multiple different pairs of languages -- even pairings the system has never been exposed to before. In a blog post Tuesday, a trio of Google researchers explain how upgrades to their recently unveiled Google Neural Machine Translation (GNMT) system pinpoint the required target language and allow for translations between different combinations of languages, sight-unseen. This means Google Translate can now translate, say, Korean into Japanese, despite having received no Korean-into-Japanese training data. An accompanying research paper was published in Cornell University's arXiv ("archive"), an online repository for scientific research papers. Last week, the Google Translate team announced it had the ability to translate whole sentences, as opposed to merely reading things as a succession of discrete individual words.

The U.S. Internet giant announced that Google Translate has been switched to a new system called Neural Machine Translation (NMT), an end-to-end learning framework that learns from millions of examples. The multilingual system is based on machine learning that provides computers with the ability to learn without being explicitly programmed. Unlike the current translation system that was adopted 10 years ago, the new system considers the entire sentence as one unit. Previous systems translated words and phrases independently within a sentence. Google said the NMT interprets entire sentences, making the translation not only sound much more like a native speaker of the language but more accurate.

In past blog posts, we discussed different models, objective functions, and hyperparameter choices that allow us to learn accurate word embeddings. However, these models are generally restricted to capture representations of words in the language they were trained on. The availability of resources, training data, and benchmarks in English leads to a disproportionate focus on the English language and a negligence of the plethora of other languages that are spoken around the world. In our globalised society, where national borders increasingly blur, where the Internet gives everyone equal access to information, it is thus imperative that we do not only seek to eliminate bias pertaining to gender or race inherent in our representations, but also aim to address our bias towards language. To remedy this and level the linguistic playing field, we would like to leverage our existing knowledge in English to equip our models with the capability to process other languages. Perfect machine translation (MT) would allow this. However, we do not need to actually translate examples, as long as we are able to project examples into a common subspace such as the one in Figure 1. Ultimately, our goal is to learn a shared embedding space between words in all languages. Equipped with such a vector space, we are able to train our models on data in any language. By projecting examples available in one language into this space, our model simultaneously obtains the capability to perform predictions in all other languages (we are glossing over some considerations here; for these, refer to this section). This is the promise of cross-lingual embeddings. Over the course of this blog post, I will give an overview of models and algorithms that have been used to come closer to this elusive goal of capturing the relations between words in multiple languages in a common embedding space. Note: While neural MT approaches implicitly learn a shared cross-lingual embedding space by optimizing for the MT objective, we will focus on models that explicitly learn cross-lingual word representations throughout this blog post. These methods generally do so at a much lower cost than MT and can be considered to be to MT what word embedding models (word2vec, GloVe, etc.) are to language modelling.

Artificial intelligence (AI) will emerge as a key marketing point in the smartphone industry next year, with the world's leading handset manufacturers scrambling to add more AI functions to their upcoming flagship devices. The move comes as big data technology, allowing devices to make human-like predictions, has been identified as a core feature for a paradigm shift in the global handset market which has stagnated in recent years. The weak growth has been attributed to toughened rivalry among existing players and a slowdown in terms of technological innovations from industry-leading handset makers ― such as Samsung Electronics and Apple ― according to observers. According to market researcher Gartner, the smartphone market has reached 90 per cent penetration in key global markets this year - including the United States, Europe and Japan ― where lifecycles for premium handsets are extended to 2.5 years on average. The trend will prolong over the next five years, added the firm.

Google Translate is getting brainier. The online translation tool recently started using a neural network to translate between some of its most popular languages – and the system is now so clever it can do this for language pairs on which it has not been explicitly trained. To do this, it seems to have created its own artificial language. Traditional machine-translation systems break sentences into words and phrases, and translate each individually. In September, Google Translate unveiled a new system that uses a neural network to work on entire sentences at once, giving it more context to figure out the best translation.

Google says its artificial intelligence has taught itself to'translate between languages that it doesn't even know' 'Zero-shot translation' can translate between languages it doesn't know Deep-learning researchers developed Google Neural Machine Translation GNMT developed algorithm that'self-teaches' it to translate languages'Zero-shot translation' can translate between languages it doesn't know GNMT developed algorithm that'self-teaches' it to translate languages Google headquarters in Menlo Park, California is seen in the above stock photo. He really is a boy's best friend! Three-year-old Reagan has... Russia is developing a mega-rocket that will transport... Frail Hugh Hefner flashes a smile while wearing his... EXCLUSIVE: Andy Cohen, 48, gets affectionate with his... Ohio state knifeman ranted about how he was'sick and tired... Trump gives Romney a SECOND secretary of state interview... 'You'll be a Man, my son!' The Duke of Westminster's son and... Case of'German Madeleine McCann' is solved after 15 years... Trump nemesis Rosie O'Donnell is slammed after speculating... Stunning new data indicates El Nino drove record highs in... He really is a boy's best friend! Russia is developing a mega-rocket that will transport... Frail Hugh Hefner flashes a smile while wearing his... EXCLUSIVE: Andy Cohen, 48, gets affectionate with his...

Machine Learning is one of the most exciting fields in the world. Every week we discover something new, something amazing, something revolutionary. It's incredible, but it can also be overwhelming. That's why we created This Week in Machine Learning! Each week we publish a curated list of Machine Learning stories as a resource to help you keep pace with all these exciting developments.

As someone who lives between two countries, I've relied heavily on machine translation for a number of years. It took me years before I could communicate casually with my in-laws (they're fluent in German and Polish, but not English), without my wife serving as an interpreter. I realize that nothing beats a skilled human translator. But who has time (or money) to hire a person to translate simple, everyday tasks? That's why I've always been amazed at computer-aided translation, specifically Google Translate.