Task 2-2

Building and maintaining the index. We aim to build the index by querying  KGs, following the research direction of [33]. Currently, KG providers need to pre-compute descriptions of their own graphs, following standard conventions. Conventions can be hard to follow, especially without strong incentives as in Google dataset Search[20]. As soon as accessibility is guaranteed (task 1), exhaustive descriptions can effectively be obtained through queries, relieving KG providers of the burden of conventions. The index will make KGs findable. Different kinds of index can be built: VOID description, Schema-index, Instance-level index, we need to ii) evaluate their maintenance and ii) determine which are tractable.

Comments are closed.