Google has faced criticism for perceived privacy breaches in the past
Google has faced criticism for perceived privacy breaches in the past

Google launches Parsey McParseface, a new algorithm inspired by the world’s most controversial boat

The new product is based on technologies with names like TensorFlow and SyntaxNet – but after being unable to pick a name, Google decided to name it after Boaty McBoatface

Andrew Griffin
Friday 13 May 2016 11:39

Google has revealed the most powerful computer for understanding the English language in the world – and called it Parsey McParseFace.

The technology, which is built on the more sensibly named TensorFlow and SyntaxNet frameworks, is a powerful tool that uses new artificial intelligence technology to be able to analyse the linguistic structure of language, and understand what each part of a sentence does to its meaning. Google is making the tool open source, so that anybody can use it for free.

But it will probably go down in history because of its silly name. Google said that the name – a reference to the controversial Boaty McBoatface – was a suggestion that came while it was trying to name the new technology, and that it didn’t have any better alternatives.

But beyond the name, Google’s new technology could change the way that artificial intelligence understands language forever.

Mr McParseface is the English language implementation of SyntaxNet, a technology that has also been put into the public domain. Both are meant as ways of better understanding the function of language, and by doing so enable computers to better be able to speak to people.

The development of the technology is key to making computers able to understand what people say to them. Since language can be so illogical, computers can have a tough time actually working out what people are telling them, but Google’s new tools attempt to overcome that.

“One of the main problems that makes parsing so challenging is that human languages show remarkable levels of ambiguity,” Google said as it launched the new tools. “It is not uncommon for moderate length sentences – say 20 or 30 words in length – to have hundreds, thousands, or even tens of thousands of possible syntactic structures.

“A natural language parser must somehow search through all of these alternatives, and find the most plausible structure given the context.”

At the moment, SyntaxNet and Parsey McParseface have a relatively limited understanding of how sentences work, far from any comparison with a human adult. But because they are built using machine learning, the algorithms will be able to train themselves and so get better at understanding as they are used.

Join our new commenting forum

Join thought-provoking conversations, follow other Independent readers and see their replies

View comments