Metadata portals offer access to Open Data, which is increasingly published on the basis of RDF vocabularies and ontologies. Individual datasets include URIs and descriptions of licenses under which the datasets are published. Licenses can be defined by attributes of the types permission, requirement and prohibition. This paper focuses on the problem of relicensing datasets that include multiple other open datasets published under different licenses. If combined datasets are republished, a chosen license must comply with the original licenses of each dataset. The approach is based on mapping licenses to deontic interpretations and combining them using an OR-composite operator. The result of the operator is used to determine matching licenses from a Knowledge Base containing license information. The approach is evaluated based on the Creative Commons License Compatibility Chart.

1. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
Relicensing Combined Datasets
Adrian Wilke • Arwa Bannoura • Axel-Cyrille Ngonga Ngomo
DICE Group
Department of Computer Science
Paderborn University, Germany
15th IEEE International Conference on Semantic Computing
Resource Track

2. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
Relicensing Combined Datasets
2
‣ Licensing Concepts & KGs
‣ ReCoDa Approach
○ Attributes to Binary Values
○ Attribute Types
○ Computation of
Composite Attributes
○ Back-Mapping
‣ Evaluation
‣ Demo & Resources

3. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
Licensing Concepts
3

4. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
License Knowledge Graphs
4
Derivative Works
Distribution
Reproduction
Attribution
Notice
Commercial Use
Distribution of derivative works
Distribution, public display, and publicly performance
Making multiple copies
Credit be given to copyright holder and/or author
Copyright and license notices be kept intact
Exercising rights for commercial purposes

5. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
ReCoDa Approach
5
1. Using Knowledge Graph: (licenses & attributes):
Mapping: License attributes to binary values
2. Semantics of attribute types:
Effects on the compatibility
3. Operator:
Generate composite attributes
4. Back-mapping of resulting attributes
🠂 List of compatible licenses

6. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
1. Mapping of Attributes to Binary Values
6

7. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
1. Mapping of Attributes to Binary Values
7

8. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
2. License Comparison Types
8
“distribution of derivative works”
Permission: 1 if not permitted
“derivative works be licensed under
the same terms or compatible
terms as the original work”
Use most restrictive license to
not to violate any original license

9. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
3. Computation of Composite Attributes
9

10. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
4. Determining Compatible Licenses
10

11. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
Evaluation: CC License Compatibility Chart
11
"Choose two works you wish to combine or remix. [...]
Use at least the most restrictive licensing of the two
(use the license most to right or down state) for the new work."
https://wiki.creativecommons.org/index.php?title=Wiki/cc_license_compatibility

12. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
Evaluation:
Creative Commons
License
Compatibility
Chart
12

13. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
612 Creative Commons licenses
Evaluation
13

14. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
612 Creative Commons licenses
Localized to languages & versions & 7 attributes → 15 data points
Evaluation
14

15. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
Demo
15
‣ Application
‣ 3 Knowledge Graphs
○ CC Chart:
8 Licenses /
7 Attributes
○ EDP: 31 Lic. / 12 Att.
○ CC: 612 Lic. / 10 Att.
‣ Webservices
○ Select Knowledge Graph
○ Select licenses
○ Get compatible licenses

16. Wilke, Bannoura, Ngonga: Relicensing Combined Datasets
Thank you for your attention!
Resources:
‣ https://github.com/dice-group/ReCoDa
○ Code
○ Knowledge Graphs
‣ https://dice-research.org/
○ Demo
○ Research Group
16