Photo Credit: Wired.com
It’s no secret that the Internet, while an incredible source of information about pretty much anything, is also filled to the brim with garbage. Anti-vaccination websites making dubious claims backed by anecdotes are on the front page of Google’s search engine results pages, and sites claiming to “news sources” are everywhere, gathering a substantial audience despite citing very few facts.
In response to this problem, Google is considering a new algorithm that grades the trustworthiness of a website or page, considering this factor when ranking it for a particular keyword. In other words, the search giant is looking at grading sites by facts, and not just links.
The Problem with the Status Quo
When you want accurate information about something, you might have read about, or heard of, what do you do? The library used to be the safest choice, as most books that made their way there often had to meet stringent standards for factuality and accuracy.
These days though, people are more likely to Google their questions.
However, just because Google leads you to what it thinks are the answers to your questions, doesn’t mean they’re the right answers. And the problem is more prevalent than you think, as Google currently factors in the number of incoming links a web page has, using it as the basis for quality and ranking on its search results. Simply put, the more sites linking to a webpage, the higher Google ranks it.
But Google’s current search engine is far from perfect, with many websites having little to no facts at all managing to rise up the rankings. After all, just because several people are linking to website, thinking the information found there is valuable, doesn’t mean the information is based on facts. Case in point: gossip websites and their loose definition of facts.
To avoid this rampant hijacking of real estate on search engine results, a team of Google researchers has created what it calls a “Knowledge-based Trust” algorithm that uses factual accuracy to rank websites instead of popularity. The challenge to making the algorithm work, is figuring out what information on a website is factual, and what isn’t.
How it Works
The system, which is still in the experimental stage, computes the number of inaccurate facts in a single web page, with a source having few inaccurate facts seen as more trustworthy. This computation results in what the team calls a Knowledge-Based Trust score. To make the algorithm work, the team uses Google’s proprietary Knowledge Vault, a massive vault of facts collected from the Internet.
The Knowledge Vault works by sniffing for information that matches a pattern Google calls “triples,” which meet the following 3 factors:
- A subject that’s a real-world entity
- A description of some characteristic of said entity
- An object that shows the value of said characteristic
Here’s a real world example:
President Obama (the real-world subject entity), is the incumbent president (predicate) of America (object).
Google’s Knowledge Vault houses billions, if not trillions, of triples gathered from the Internet, with the Knowledge-based Trust algorithm using the vault to determine whether facts on a web page are true or not.
This reinforces the value of quality, factual data for website owners about the products and services they are offering.
Need help with content? Contact the team at Enform.