Knowledge bases like the TAC KBP graph are noisy. We want to know which algorithm is most suitable for label acquisition?
There are existing algorithms for KBC, but they can only be applied to attribute propagation through a convoluted step of reducing the Vertex Labeling problem to Knowledge Base Completion.
However, what we want to know is that what are the existing frameworks for Knowledge Base Completion and Vertex Labeling? And which of these are going to be reasonable baselines?