How Can We Complete a Knowledge Graph?
Details
A Knowledge Graph is as valuable as the insights we can derive from it. So, what do we do when our Knowledge Graph doesn’t contain the answers? We need to complete it.
We know that Grakn’s logical reasoner can help us to deduce insights. However, when our answers can’t be deduced we need to turn to statistical methods to infer new facts - making predictions inductively, by example. This could be relations, attributes or even rules.
In this talk, we will delve into the advanced graph learning systems that we can construct and use on top of Grakn to create intelligent systems. This is the core of the research that we conduct at Grakn Labs - all of which is made available in KGLIB.