| Abstract |
In the first decade of its existence, datasets from the LOD Cloud have
been stored, indexed, and exposed in triple stores that implement
(some subset of) the SPARQL standard. While these triple stores are
useful for some purposes, in recent years we have seen an increasing
number of Linked Data use cases that do not work well with -- and that
may ultimately not depend on -- the traditional SPARQL paradigm.
Examples include big data processing, graph navigation, cost-effective
publishing, machine learning, and empirical semantics.
In this talk I will give a brief overview of where HDT and ClioPatria
development is heading today, and what are the main use cases that it
seeks to enable. I will specifically delve into the use case of
empirical semantics, where we wish to determine the meaning of
expressions empirically (rather than by analytic means). I will focus
on the challenges that we face when studying semantics empirically,
and how we can overcome these challenges by pushing forward the
'semantic frontier'. |