Has anyone wondered how Google suggests related sentences whenever someone searches something on Google? How do you think Google came up with it? Does it leave you wondering how they manage to program its interface? Well, Google has an answer to those questions.
According to The Verge, Google recently released its software tool SyntaxNet and component of it called Parsey McParseface.These are the tools that Google uses to understand natural language when people type in something into the box or speak to Google Now. SyntaxNet is the overall syntactic parser. According to Google, the particular artificial intelligence is like a trained human linguist because it correctly identifies grammatical features and building blocks, which includes the subject, object and verbs. Google even claimed that it is 94 percent accurate.
According to The Wall Street Journal, Parsey McParseface is a computer program that helps machine understand written English. In addition to the post, Google is offering the code SyntaxNet free for anyone to distribute or modify.
The Mirror stated that the said technology is so powerful because it is able to analyze the linguistic structure of language from simple to complex. Additionally, Google stated that Parsey McParseface uses artificial intelligence to consider multiple interpretations of a sentence. It gives plausible score and selects which one has the highest. Google's benchmarking test on it shows that the artificial intelligence beats the company's own record. It achieved 94 percent accuracy, claiming that it is already approaching near human.
How did Google come up with the term 'Parsey McParseface'? According to The Mirror, the name is a tribute to a competition launched by the Natural Environment Research Council to name a polar research ship. The term "parsing" comes from a Latin word "pars," which means part of speech. For more information about its lexical matter, read here.
