Knowledge Graphs are an information system design pattern that is gaining a lot of attention. Originally developed by Google and other large companies, the techniques are now spreading out to mere mortals. In this talk we will discuss what is a knowledge graph (and what it isn't), then look at technology for building a knowledge graph using W3C standards.
5. What is a Knowledge Graph?
https://2019.semantics.cc/knowledge-graphs-are-rise
● A Knowledge Graph represents a knowledge domain
● It represents knowledge as a graph
● a network of nodes and links
● not tables of rows and columns
● It represents facts (data) and models (metadata) in the same way
6. WHAT IS A KNOWLEDGE GRAPH?
“A knowledge graph acquires and integrates information into an
ontology and applies a reasoner to derive new knowledge.”
Towards a Definition of Knowledge Graphs
SEMANTICS 2016: Posters and Demos Track
7. WHAT IS A KNOWLEDGE GRAPH?
“A knowledge graph acquires and integrates information into an
ontology and applies a reasoner to derive new knowledge.”
Towards a Definition of Knowledge Graphs
SEMANTICS 2016: Posters and Demos Track
8. Reality
Data
“Integrates” => can be many sources;data come from other places
“Ontology” => organised
“Reasoner”, “machine learning” => programs that make the graph more useful
9. Another take . . .
https://medium.com/@dmccreary/knowledge-graphs-the-third-era-of-computing-a8106f343450
12. Knowledge Graph
● An approach, not a specific technology
● Integrates
○ Many sources
○ Variety of sources
● Refines the data
○ Puts some level of organisation on the data
○ Makes connections across sources
● Answer questions
○ Supplies an API or a UI for delivery of organised data
○ “Publishes data” : not a single planned, fixed usage.
13. Now:
● General term for a graph of data produced from many sources.
● Background knowledge for AI and ML
17. While the term “Knowledge Graph” is relatively new
(Google 2012), the concept of “representing knowledge
as a set of relations between entities — forming a
“graph” — has been around for much longer.
https://dzone.com/articles/my-list-of-7-great-2018-advancements-in-enterprise
19. Why Graph Databases
● “Schemaless” and Publishing
● Data not completely regular … or clean … or cleanable
● “Connections”
● Traversal
20. Two kinds
RDF
● W3C Standards
● From 1999
● Focus:
○ Data Publishing
○ Data Integration
○ Web scale
○ Modelling
○ Rules
● Query language: SPARQL
Property Graphs
● A general style, currently
converging
● Focus:
○ Data capture
○ Analytics
○ Large datasets
○ Data structure
● Query languages: Gremlin,
OpenCypher
36. Summary
Knowledge Graphs for data integration
Graph databases to contain variable shape data
RDF databases for standard-based data exchange, modelling and query.
Open source systems to build Knowledge Graphs.