Facebook's engineering team has published yet another round of nitty-gritty details holding up the Social Graph.
This time, Open Graph third-party developers and members of the world's largest social network alike are being informed of how the dataset connects them to other people, places and pages.
Referred to as the "Entity Graph," the infrastructure is designed "to bring structure and meaning to data."
That's not an easy task given the sheer amount of content floating around the network.
For one, Facebook reiterated that it now retains more than one billion monthly active users -- 600 million of whom log in each day.
That translates into the more than 100 billion pieces of content (and later connections) that form the Entity Graph.
See also: Understanding Unicorn: A deep dive into Facebook's Graph Search | Facebook translates natural language interface under Graph Search | Facebook intros changes to simplify Open Graph setup for developers
Eric Sun, a software engineer on the entities team, outlined some of the challenges in a blog post on Thursday.
Some of the factors that go into forming those connections for the end user range include removing junk Pages, matching Wikipedia data, and manual labeling for the largest nodes/categories (i.e. movies).
Sun explained further:
Every day, the entities graph grows with new nodes and connections at a pace far greater than our small team can manually inspect. Therefore, we concentrate on building multi-pronged systems that are scalable and that will improve the graph over time. Current efforts are focused on cleaning up the long tail via back-end machine learning as well as expanding crowdsourcing efforts to allow all of our users to contribute corrections and additions to our entity pages.
Background information and more technical information about the path developers have taken to reach this point in the Entity Graph are available on The Facebook Engineering Blog now.
Images via The Facebook Engineering Blog