Tracking Data Sources of Fused Entities in Law Enforcement Graphs
0 likes84 views
This document discusses modeling considerations for law enforcement graphs. It addresses tracking data sources, modeling fused entities from multiple sources, and representing information credibility ratings. Key challenges include how to model data sources and hierarchies, represent fused entities and related facts, and handle updates given different information sources. The document provides examples of modeling sources as nodes, labels, and properties and discusses tradeoffs between materializing or linking fused entities and facts.
Tracking Data Sources of Fused Entities in Law Enforcement Graphs
- 1. © 2022 Neo4j, Inc. All rights reserved. © 2022 Neo4j, Inc. All rights reserved. Tracking Data Sources of Fused Entities in Law Enforcement graphs Luanne Misquitta, VP of Engineering, GraphAware
- 2. © 2022 Neo4j, Inc. All rights reserved. Graphs and Law Enforcement 2
- 3. © 2022 Neo4j, Inc. All rights reserved. 3 Connect the dots
- 4. © 2022 Neo4j, Inc. All rights reserved. 4 Connect the dots
- 5. © 2022 Neo4j, Inc. All rights reserved. 5 Connect the dots
- 6. © 2022 Neo4j, Inc. All rights reserved. 6 Connect the dots
- 7. © 2022 Neo4j, Inc. All rights reserved. 7 Connect the dots
- 8. © 2022 Neo4j, Inc. All rights reserved. Outline • Key properties of law enforcement graphs ◦ Data sources ◦ Data quality and reliability ◦ Fused entities • Modelling considerations 8
- 9. © 2022 Neo4j, Inc. All rights reserved. POLE 9 Person Location Event Object
- 10. © 2022 Neo4j, Inc. All rights reserved. Graph model 10
- 11. © 2022 Neo4j, Inc. All rights reserved. The graph 11
- 12. © 2022 Neo4j, Inc. All rights reserved. Disconnected events 12
- 13. © 2022 Neo4j, Inc. All rights reserved. 13 Data Provenance • The source of a piece of information and the process that led to its existence • Used to form assessments about the quality, reliability and trustworthiness of the information at hand
- 14. © 2022 Neo4j, Inc. All rights reserved. 14 The INTs • Intelligence Collection disciplines • Hierarchical • Examples are SIGINT, OSINT, MASINT, GEOINT
- 15. © 2022 Neo4j, Inc. All rights reserved. Local Crime data 15
- 16. © 2022 Neo4j, Inc. All rights reserved. Bring in Forensics 16
- 17. © 2022 Neo4j, Inc. All rights reserved. Modelling sources- use cases • Do analysts want to only see the source that provided the information or do more with it? • Do analysts expect to traverse the graph within a particular source? • Are sources attached only to nodes or also to relationships? • Are sources hierarchical and does querying a source at a higher level imply including all child sources? • Does access control need to be set up to grant/deny access to certain sources? • Security classification levels by source 17
- 18. © 2022 Neo4j, Inc. All rights reserved. Modelling sources- as a node • Hierarchies are modelled naturally • Source based access control handled through the presence or absence of a path • Sources on relationships do not fit into this model • Supernode problem 18
- 19. © 2022 Neo4j, Inc. All rights reserved. Modelling sources- as a label • For nodes, not relationships • Hierarchy can be replicated through labels • Label based access control 19
- 20. © 2022 Neo4j, Inc. All rights reserved. Modelling sources- as a property • Both nodes and relationships • Source hierarchy maintained separately, parameters passed via queries • Access control on property values not supported natively • Test performance when filtering on sources when traversing 20
- 21. © 2022 Neo4j, Inc. All rights reserved. Layering for analysis 21
- 22. © 2022 Neo4j, Inc. All rights reserved. Information and Source Ratings Source Reliability is the degree of confidence assigned to the source with respect to its capacity to provide valuable and accurate information. https://en.wikipedia.org/wiki/Intelligence_source_and_information_reliability 22
- 23. © 2022 Neo4j, Inc. All rights reserved. Information and Source Ratings Information Credibility is the confidence assigned to the information with respect to its veracity. https://en.wikipedia.org/wiki/Intelligence_source_and_information_reliability 23
- 24. © 2022 Neo4j, Inc. All rights reserved. Missing Information and Source Ratings 24
- 25. © 2022 Neo4j, Inc. All rights reserved. Drawing a different conclusion 25
- 26. © 2022 Neo4j, Inc. All rights reserved. Modelling Information and Source ratings • A property for each on nodes and relationships • Source hierarchy defaults ◦ Compose for display purposes ◦ Copy to filter during navigation 26
- 27. © 2022 Neo4j, Inc. All rights reserved. Same entity, different sources 27
- 28. © 2022 Neo4j, Inc. All rights reserved. Hypothesis links 28
- 29. © 2022 Neo4j, Inc. All rights reserved. Using node representations 29
- 30. © 2022 Neo4j, Inc. All rights reserved. Grouping 30
- 31. © 2022 Neo4j, Inc. All rights reserved. Create a fused entity • Create a new entity which fuses information from facts determined to be the same • Determine the source and confidence in the new fused entity 31
- 32. © 2022 Neo4j, Inc. All rights reserved. Copy relationships from the facts • The fused entity stands in for the underlying facts • Relationships are transferred from the facts to the fused entity 32
- 33. © 2022 Neo4j, Inc. All rights reserved. Path finding 33
- 34. © 2022 Neo4j, Inc. All rights reserved. What about the facts? • Create a FACT relationship from the fused entity to the facts • Allows analysts to drill in and view individual facts • Could affect node expansion and path finding • Clutters visual exploratory analysis 34
- 35. © 2022 Neo4j, Inc. All rights reserved. Offloading facts to another graph Create the fused entity in the fact graph (shell entity) 35 Main graph Main graph Fact graph
- 36. © 2022 Neo4j, Inc. All rights reserved. Offloading facts to another graph Recreate each fact in the fact graph and create the FACT relationship to the fused entity 36 Fact graph
- 37. © 2022 Neo4j, Inc. All rights reserved. Offloading facts to another graph Copy each entity (shell) that the fact in the main graph is linked to and link them to the facts in the fact graph 37 Fact graph
- 38. © 2022 Neo4j, Inc. All rights reserved. Offloading facts to another graph Remove the facts from the main graph 38 Main graph
- 39. © 2022 Neo4j, Inc. All rights reserved. Loading facts on demand 39
- 40. © 2022 Neo4j, Inc. All rights reserved. Challenges with updates to fused entities • Do fused entities automatically get updated when the facts are updated? • Materialised vs shell fused entities • Information and source ratings may no longer be valid • Conflicts need to be handled 40
- 41. © 2022 Neo4j, Inc. All rights reserved. Takeaways • Data sources ◦ Consider use cases before modelling- single/multivalued, nodes/relationships, hierarchy, access control ◦ Performance testing if filtering during traversals • Importance of Information and Source Ratings • Fused entities ◦ Via links or nodes ◦ Materialise or compose at query time ◦ Representing facts- same or different graph ◦ Consider flavours of fused entities if attribute level sources are to be tracked • Fact update challenges ◦ System of record ◦ Validity of information and source rating of the fused entity post update ◦ Handling conflicts 41
- 42. © 2022 Neo4j, Inc. All rights reserved. © 2022 Neo4j, Inc. All rights reserved. 42 Thank you! Or, contact us at info@graphaware.com Visit GraphAware’s booth to explore the graph