Knowledge graphs (KGs) emerged in 2011 with the publicizing of the Google Knowledge Graph as ‘things not strings’.
In a few turbulent years, and many publications and applications later, KGs have entered mainstream terminology.
In its 2019 hype cycle, Gartner put KGs as halfway to the ‘peak of inflated expectations’. The very next year, KGs did not even make the list: instead, ‘ontologies and graphs’ are nearly at the ‘trough of disillusionment’.
These findings suggest both terminological confusion and a misunderstanding of the past, present and potential future of KGs. This talk will serve a dual purpose.
On the one hand, it will take a decade-long perspective of KGs, and place it in context of other technologies and innovations in the artificial intelligence community.
On the other hand, it will present an ambitious and future-facing ‘moonshot’ perspective on KGs.
It will also offer some evidence-based opinions on how best to navigate this frontier.