Googly McSearchface has released SyntaxNet, “an open-source neural network framework” and an open source tool for parsing the English language called Parsey McParseface.
The company hopes that Codey McCoderfaces will put the two tools to work doing Natural Language Understanding (NLU), the art of helping computers to understand spoken language.
The Alphabet subsidiary’s Researchy McResearchface branch thinks that SynatxNet is just the tool to help Parsey McParseface do all the heavy lifting needed to process language. Googly McSearchface senior staff research scientist Slav Petrov says a neural network is needed because “human languages show remarkable levels of ambiguity.”
“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,” Petrov writes offering the sentence “Alice drove down the street in her car” as an example of the challenges parsers face because one interpretation is that Alice drives down a street while driving her car. The other is that the street in question is inside Alice's car.
Humans quickly understand the second interpretation is nonsensical. Machines can take things rather more literally, so giving them a neural network to help them crunch parsing permutations helps things along.
Parsey McParseface can apparently get this stuff right 94 per cent of the time, making it rather better than comparable code.
Both tools are yours for the downloading at Gitty McGitHub. ®
Bootnote: Boaty McBoatface was fun, but enough is enough. Or after this story. surely it is?