Successfully reported this slideshow.
Your SlideShare is downloading. ×

Government GraphSummit: Digital Transformation with Graphs, Ontology and ML Ops

Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Loading in …3
×

Check these out next

1 of 40 Ad

Government GraphSummit: Digital Transformation with Graphs, Ontology and ML Ops

Download to read offline

Steven Scott, Cognitive Software Engineer and Zach Tretter, Data Scientist, Digital Innovation, Northrop Grumman

Digital Transformation is a core focus of Northrop Grumman. The Space Systems Digital Innovation Team combines Graph, Ontology-Driven Design (ODD), and ML Ops to solve the “drowning in data but starving for knowledge” problem. ODD operates on the premise that data integration should be governed by knowledge. In this paradigm, domain knowledge is modeled ontologically, which kills two birds with one stone. In this talk, we demonstrate how Northrop Grumman uses Neo4j graph databases to realize ODD pipelines that generate knowledge graphs from which analytical methods and ML Ops transform data and domain knowledge into customer value. The Digital Innovation Team applies this pattern to projects encompassing Fleet Management, Master Data Management, Digital Twins, Schedule Optimization, and Prognostics.

Steven Scott, Cognitive Software Engineer and Zach Tretter, Data Scientist, Digital Innovation, Northrop Grumman

Digital Transformation is a core focus of Northrop Grumman. The Space Systems Digital Innovation Team combines Graph, Ontology-Driven Design (ODD), and ML Ops to solve the “drowning in data but starving for knowledge” problem. ODD operates on the premise that data integration should be governed by knowledge. In this paradigm, domain knowledge is modeled ontologically, which kills two birds with one stone. In this talk, we demonstrate how Northrop Grumman uses Neo4j graph databases to realize ODD pipelines that generate knowledge graphs from which analytical methods and ML Ops transform data and domain knowledge into customer value. The Digital Innovation Team applies this pattern to projects encompassing Fleet Management, Master Data Management, Digital Twins, Schedule Optimization, and Prognostics.

Advertisement
Advertisement

More Related Content

More from Neo4j (20)

Recently uploaded (20)

Advertisement

Government GraphSummit: Digital Transformation with Graphs, Ontology and ML Ops

  1. 1. Digital Transformation with Graphs, Ontology, and ML Ops at Northrop Grumman 7 December 2022 Data Scientist Zachary Tretter 1 Steven Scott Cognitive Software Engineer Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  2. 2. • Data Challenges for Digital Transformation • Methodology – Ontology Driven Design • Applications of Ontology Driven Design Outline 2 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  3. 3. About Northrop Grumman 3 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  4. 4. Data Problems Siloed Dirty Incomplete Rigid Obscure Difficult to Understand 4 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  5. 5. 5 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  6. 6. 6 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  7. 7. Drowning in data but starving for knowledge 7 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  8. 8. 8 Disparate Data + SME Knowledge Model = Context-Rich Knowledge Graph Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  9. 9. Ontology-Driven Design 9 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  10. 10. “Software should be governed by knowledge, not by code” - The ODD Design Philosophy 10 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  11. 11. 11 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  12. 12. • Identify what things exist • Identify relationships between those things • Disambiguate between conflated terms • Map Data Sources to Graph Structures to Questions How to Build an Ontology purchases actor actor relationship axiom User shoes User purchases shoes “Our users generally purchase shoes from us every 11 months” 12
  13. 13. 13 Thou Shalt A… Thou Shalt B… Thou Shalt C… Thou Shalt D… Thou Shalt E… IMPERATIVE DECLARATIVE Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  14. 14. 14 Thou Shalt A… Thou Shalt B… Thou Shalt C… Thou Shalt D… Thou Shalt E… IMPERATIVE DECLARATIVE ASYMMETRIC MAPPING Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  15. 15. 15 CODE F C B E D A ONTOLOGY DECLARATIVE IMPERATIVE SYMMETRICAL MAPPING! Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  16. 16. 16 F C B E D A IMPERATIVE DECLARATIVE SYMMETRICAL MAPPING! Thou Shalt A… Thou Shalt B… Thou Shalt C… Thou Shalt D… Thou Shalt E… Requirements Ontology Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  17. 17. 17 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  18. 18. ure dolor in reprehenderit 101010010101010 y”},{obj:”property”},{obj:”property”},{obj:”property”},{obj:”property”}, {obj:”property”},{obj:”property”},{obj:”property”},{obj:”property”},{obj:”property”},{obj: 18 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  19. 19. 19 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  20. 20. 20 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  21. 21. 21 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  22. 22. 22 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  23. 23. 23 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  24. 24. 24 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  25. 25. 25 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  26. 26. 26 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  27. 27. Build Knowledge Models 27 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  28. 28. Build Knowledge Models 28 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  29. 29. Derive Mapping Pipelines 29 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  30. 30. Derive Mapping Pipelines 30 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  31. 31. Analyze Smart Data 31 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  32. 32. Our Graph and Ontology-Driven Design Projects 32 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  33. 33. Digital Twins 33 Compilation of Sensor Driven observations informing current state The ideal design created by engineers and planners prior to realization Ideal State Twins Actual State Twins Projected future state based on analytics Potential State Twins Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  34. 34. Digital Twinning • Buildings, rooms, equipment, employees • View seating, conference room use • Planning changes 34 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  35. 35. Fleet Management Maintenance Data Sensor Data System Understanding Reliability Sustainability Supply Chain Supportability Optimized Solution Budget Desired Availability Manpower Constraints Customer Driven Cascading Predictive Analysis 35 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  36. 36. Master Data Management 36 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  37. 37. Master Data Management 37 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  38. 38. Discourse Graphs 38 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  39. 39. Questions? 39 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman
  40. 40. 40 Approved for Public Release: NG22-2209 © 2022, Northrop Grumman

×