Diese Präsentation wurde erfolgreich gemeldet.
Wir verwenden Ihre LinkedIn Profilangaben und Informationen zu Ihren Aktivitäten, um Anzeigen zu personalisieren und Ihnen relevantere Inhalte anzuzeigen. Sie können Ihre Anzeigeneinstellungen jederzeit ändern.

Algebraic Property Graphs

63 Aufrufe

Veröffentlicht am

Slides about contents of the paper of Joshua Shinavier and Ryan Wisnesky. Presented by Adrian Wilke at the DICE (Data Science Group) colloquium.

Veröffentlicht in: Wissenschaft
  • Als Erste(r) kommentieren

  • Gehören Sie zu den Ersten, denen das gefällt!

Algebraic Property Graphs

  1. 1. Algebraic Property Graphs Joshua Shinavier Ryan Wisnesky https://arxiv.org/pdf/1909.04881.pdf DICE colloquium presentation Adrian Wilke 2019-10-25 https://dice-research.org/
  2. 2. Authors 2 Joshua Shinavier Uber Technologies One of the three co-founders of TinkerPop [JS] image: https://on.mktw.net/2MbIZTW Ryan Wisnesky Conexus AI “[...] algebraic property graphs as a simple but mathematically rigorous bridge between graph and non-graph data models [...]” B.S. and M.S. degrees in mathematics and computer science from Stanford University and a Ph.D. in computer science from Harvard University, where he studied the design and implementation of provably correct software systems. Previously, he was a postdoctoral associate in the MIT department of mathematics, where he developed the categorical query language, CQL. He currently leads open-source and commercial development of CQL as CTO of Conexus AI. https://www.wisnesky.net APG listed as case study for Categorical Query Language (CQL) at categoricaldata.net/papers
  3. 3. Contents 3 15 text pages 7.6 pages appendix 2.3 pages toc & ref
  4. 4. Contents 4 “main contribution” motivation short introduction
  5. 5. Motivation: Property graphs ⇄ RDF 5 slide source: [CDL]
  6. 6. Motivation: Property graphs ⇄ RDF 6 slide source: [CDL]
  7. 7. Motivation: Property graphs 7 also: missing semantics in several implementations slide source: [CDL]
  8. 8. Motivation: Hypergraphs 8 slide source: [CDL]
  9. 9. Category theory 9
  10. 10. Category theory 10 “In mathematics, abstract nonsense [...] are terms used by mathematicians to describe abstract methods related to category theory.” https://en.wikipedia.org/w/index.php?title=Abstract_nonsense&oldid=881291659
  11. 11. Category theory 11
  12. 12. Category theory: Morphisms and composition 12
  13. 13. example 2 source: [EC] Category theory: Morphisms and composition 13
  14. 14. Category theory: Functors 14
  15. 15. Algebraic Property Graphs Definition 15
  16. 16. AGP: - P - t ∈ P primitive types - V(t) primitive values - four sets - four functions - equation 16 AGP: definition
  17. 17. 17 AGP: definition labels: think of classes in OOP
  18. 18. 18 AGP: definition t1 ⨉ t2 e.g. for edges edge(nodeA, nodeB) think of types in OOP
  19. 19. 19 AGP: definition value used for vertex labels elements: think of objects in OOP
  20. 20. AGP: definition 20 elements: think of objects in OOP
  21. 21. AGP: definition 21 labels: think of classes in OOP lambda like label elements: think of objects in OOP
  22. 22. AGP: definition 22 sigma 🠖 schema tau 🠖 type G(𝓣): types of G G(𝓣): types of G
  23. 23. AGP: definition 23
  24. 24. AGP: definition 24 Composition g ○ f g( f(x) ) Image: https://en.wikipedia.org/wiki/Morphism
  25. 25. Taxonomy & examples 25
  26. 26. Taxonomy: Vertices 26
  27. 27. Taxonomy: Vertices 27
  28. 28. Taxonomy: Edges 28 t1 ⨉ t2 e.g. for edges edge(nodeA, nodeB) We already saw this.
  29. 29. Taxonomy: Edges 29
  30. 30. Taxonomy: Properties 30
  31. 31. Taxonomy: Properties 31
  32. 32. Migrating AGPs 32
  33. 33. Migrating AGPs 33 obviously not.
  34. 34. RDF ⇄ AGP 34 free category is the category that results from freely concatenating arrows together, whenever the target of one arrow is the source of the next.
  35. 35. RDF ⇄ AGP 35
  36. 36. Summary 36 “We propose algebraic property graphs as a simple but mathematically rigorous bridge between graph and non-graph data models.”
  37. 37. Summary 37 “main contribution” motivation short introduction
  38. 38. For further research 38 ● Java implementation https://github.com/CategoricalData/CQL/tree/master/src/main/java/catdata/apg ● Coq proofs (paper footnote 9) https://www.categoricaldata.net/APG.v ● [CDL] Joshua Shinavier: In Search of the Universal Data Model (Connected Data London 2019) https://youtu.be/telyBQCuq70 ● Joshua Shinavier: Algebraic Property Graphs (GQL Community Update) https://www.slideshare.net/joshsh/algebraic-property-graphs-gql-community-update-oct-9-2019 ● Research on CQL https://www.categoricaldata.net/papers Related to RDF: “Formal Modelling and Application of Graph Transformations in the Resource Description Framework” (doctoral dissertation) ● [JS] Joshua Shinavier: Position statement for the W3C Workshop on Graph Data, 2019 https://www.w3.org/Data/events/data-ws-2019/assets/position/Joshua%20Shinavier.pdf ● [EC] Category Theory in Life - Eugenia Cheng https://youtu.be/ho7oagHeqNc

×