Google’s Knowledge Vault Helps Rank Sites By Accuracy

Just the facts, ma'am
A collection of switches and routers that keep Google's data centers in contact. Google

Imagine you want to figure out how accurate something is. How would you go about it? Well, you might go to the library, find the relevant encyclopedia article, and read it. But let’s be honest: in this day and age, you’ll probably just type your question into Google.

Just because Google tells you something, though, doesn’t mean it’s true. A team of Google researchers has devised a “Knowledge-based Trust” algorithm that attempts to rank websites based on their factual accuracy, rather than their popularity. In order to do that, however, it needs to be able to figure out what’s factual and what isn’t.

Google’s Knowledge Vault tries to find information that falls into a pattern of what Google calls “triples,” which are made up of three factors: a subject that’s a real-world entity, a predicate that describe some attribute of that entity, and an object that is the value of the attribute. For example, that President Obama (subject) is the president (predicate) of the United States (object).

The Knowledge Vault contains billions of those triples from across the web. And the Knowledge-based Trust algorithm uses a complicated multi-layer approach to weigh whether or not particular facts are true.

If the system works as well as hoped, Google might be able to rank sites based on just how factual they are, which is good for everything from fact-checking politicians to writing that research paper.

[via New Scientist]