Forget Boaty McBoatface! Google's AI language software is called Parsey McParseface in honour of the failed bid to name a research boat

It was a public engagement exercise that turned into a national row after a public poll to name a new British research vessel saw the name Boaty McBoatface get the most support.

But Google has had a cheeky dig at the humourless officials in the British government who objected to the name by calling its latest artificial intelligence software Parsey McParseface.

The technology firm, which has just become the world's largest company, has released software that uses sophisticated machine learning to analyse the linguistic structure of language.

Parsey McParseface uses sophisticated machine learning algorithms to analyse linguistic structure of sentences to help computers to understand how human languages are constructed. An example of of how the software decodes the sentence 'I booked a ticket to Google' is pictured

It forms part of an open-source neural network aimed at training computers to understand how human languages are put together so they can be processed.

It announced the English version of this system is called Parsey McParseface in a blog that outlined how it hopes the software will transform artificial intelligence.

The company said the algorithms would allowed computers to deal with ambiguity in language.

It said: 'Humans do a remarkable job of dealing with ambiguity, almost to the point where the problem is unnoticeable. 

'The challenge is for computers to do the same.

'Parsey McParseface recovers individual dependencies between words with over 94 per cent accuracy, beating our own previous state-of-the-art results, which were already better than any previous approach.

'While there are no explicit studies in the literature about human performance, we know from our in-house annotation projects that linguists trained for this task agree in 96 to 97 per cent of the cases.

'This suggests we are approaching human performance - but only on well-formed text.

'Sentences drawn from the web are a lot harder to analyse, as we learned from the Google WebTreebank, released in 2011.

'Parsey McParseface achieves just over 90 per cent of parse accuracy on this dataset.'

The Parsey McParseface AI model is part of a larger system called SyntaxNet which parses the functional role of each word in a sentence.

Parsey McParseface was able to easily identify the meaning of this sentence above by identifying that Alice was the subject of the verb saw and Bob is its direct object

In a simple example given by Google, the sentence Alice saw Bob contains two nouns and a verb, where Alice is the subject of saw and Bob is its direct object.

When presented with this sentence, Parsey McParseface was able to analyse this correctly.

But in more complex examples, where sentences can be 20 to 30 words in length, there can be hundreds or even thousands of possible syntactic structures.

The sentence: 'Alice drove down the street in her car' has at least two possible dependency parses.

The correct interpretation is that Alice is driving in her car, but it can also be interpreted as the street is in Alice's car.

For humans unravelling this is a relatively simple task as one is far more plausible than the other, but for a computer, it is less obvious.

For more complex sentences there can be several possible interpretations. In the example above there are at least two. The one above shows that Alice is driving her car

In the second, rather implausible interpretation (pictured) of the sentence above, the street is in Alice's car

SyntaxNet and Parsey McParseface processes sentences left to right so dependencies are added incrementally and given scores based on the plausibility.

It then uses this to work out what the correct interpretation of a sentence is.

Google has not explicitly said why it chose the name Parsey McParseface but it comes just days after MPs at the House of Commons Science and Technology Committee held an inquiry into the naming of a new research vessel by the National Environment Research Council.

A public vote was held to decide on the name with Boaty McBoatface coming top.

However, government ministers and officials were not impressed and instead the vessel was named after the broadcaster Sir David Attenborough with a submersible being called Boaty McBoatface.  

Google said the key to helping AI deal with human language is getting it to cope with ambiguity. This is something humans do well but machines are less good at (analysis of a complex sentence is pictured)

The UK's National Environment Research Council held a public poll to name its new research vessel (pictured) with the name Boaty McBoatface coming out as the favourite. Google's choice of name for its new software is thought to be a cheeky jibe at the row that ensued