Google is in the process of building a successor to its Knowledge Graph, and it will be the “largest store of knowledge in human history,” according to New Scientist.
The system, known as Knowledge Vault, automatically gathers facts from around the web and uses machine learning to turn the data into useful information. No human input is necessary.
This in is contrast to Knowledge Graph, which relies on a human-curated set of third-party sources for its information. This is then displayed in the characteristic Knowledge Graph info box when certain search queries are entered.
Currently, Knowledge Graph is bigger than Knowledge Vault, but with the potential speed and scope of machine learning, the new system is likely to surpass its predecessor soon – and have applications far beyond search.
1.6 billion facts have been gathered so far, of which 271 million are considered to be ‘confident facts’. Google’s algorithms rate the accuracy of a fact by cross-referencing new information with facts it already knows.