=Paper=
{{Paper
|id=Vol-3020/KR4L_paper_2
|storemode=property
|title=Knowledge Graphs: Meeting Point of Knowledge Representation and Representation Learning
|pdfUrl=https://ceur-ws.org/Vol-3020/KR4L_paper_2.pdf
|volume=Vol-3020
|authors=Emanuel Sallinger
|dblpUrl=https://dblp.org/rec/conf/ecai/Sallinger20
}}
==Knowledge Graphs: Meeting Point of Knowledge Representation and Representation Learning==
<pdf width="1500px">https://ceur-ws.org/Vol-3020/KR4L_paper_2.pdf</pdf>
<pre>
Knowledge Graphs: Meeting Point of Knowledge
 Representation and Representation Learning

                                         Emanuel Sallinger

                               TU Wien and University of Oxford




Knowledge Graphs (KGs) in recent years have become a melting pot of Artificial
Intelligence (AI) technologies. Two at first glance very different types of AI play
a particularly important role here: Knowledge Representation and Reasoning
(KRR) and Representation Learning (RL). The area of KRR is often associated
with traditional AI techniques, while RL is associated with what is typically
summarized as machine learning (ML).
Yet, while in Knowledge Graphs both types of technologies come together, there
currently is a perceived disconnect between the areas of RL and KRR. Most
of the research is currently concentrated on one area or the other, yet arguably
representation learning is central to making use of knowledge representation and
reasoning techniques in modern, scalable AI applications.
This plenary talk opens the International Workshop on Knowledge Represen-
tations and Representation Learning (KR4L) 2020, part of the European Con-
ference on Artificial Intelligence (ECAI) 2020. In this plenary talk, we give an
overview of the erea of Knowledge Graphs with a particular focus on KR and
RL. In particular it is divided into four parts:

   Modern Knowledge Graphs
 – A motivation for the use of modern Knowledge Graphs.
 – Financial Knowledge Graphs as a particularly interesting area.

   A Melting Pot of Technologies
 – Types of technologies meeting in Knowledge Graphs
   (data exchange and integration, data wrangling, graph databases, business
   intelligence tools, reasoners, machine learning frameworks, etc.).

   A Meeting Point of Research
 – Areas of the narrower and wider research fields related to Knowledge Graphs,
   and in particular KRR and RL.
 – The evolution of these areas in the last few years.



  Copyright © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0
  International (CC BY 4.0).
2         E. Sallinger




      Perspectives and Future
    – We are going to consider multiple perspectives on Knowledge Graphs, in-
      cluding the layered perspective that looks at KGs as representation tools,
      management systems and application services.
    – The future of KRR and RL in KGs.

We also bridge the six accepted regular papers, one invited paper and two pan-
els of the International Workshop on Knowledge Representations and Represen-
tation Learning (KR4L) 2020, spanning many aspects of KRR and RL. This
includes the use of RL and ML techniques for KRR:

    – Traversing Knowledge Graphs with Good Old (and New) Joins
    – An Evolutionary Algorithm for Rule Learning over Knowledge Graphs

The use of KRR in RL and ML:

    – Cluster Discovery from Sensor Data Incorporating Expert Knowledge
    – The Effect of Rule Injection in Leakage Free Datasets
    – A Performance Strategy: Multiple Slices of a KGE Model in Low Dimensions

And finally those where KRR and RL/ML techniques are used jointly:

    – Weaving Enterprise Knowledge Graphs:
      The Case of Company Ownership Graphs (Invited Paper)
    – Blockchains as Knowledge Graphs – Blockchains for Knowledge Graphs
      (Vision Paper)

We also give an overview of the two panels:

    – Emerging Topics in Academia and Industry
      Lead by Luigi Bellomarini
      (Deputy Director of IT Research, Central Bank of Italy)

    – Future Directions - Looking Ahead
      Lead by Sahar Vahdati and Mojtaba Nayyeri
      (InfAI and University of Bonn)


Acknowledgements. This work was supported by the WWTF (Vienna Science and
Technology Fund) grant VRG18-013, the EPSRC grant EP/M025268/1, the EU Hori-
zon 2020 grant 809965, and the Raison Data Royal Society grant.

</pre>