Provenance in Data Science: From Data Models to Context-Aware Knowledge Graphs

Preface
Contents
About the Editors
1 The Evolution of Context-Aware RDF Knowledge Graphs
	1.1 Introduction to RDF Data Provenance
	1.2 Extensions of the Standard RDF Data Model
	1.3 Extensions of RDFS and OWL
	1.4 Alternate Data Models and NoRDF Knowledge Representations
	1.5 RDF Graph Decomposition
	1.6 Encapsulating Provenance with RDF Triples
	1.7 Capturing Context: Graph-Based Approaches
	1.8 Utilizing Vocabularies and Ontologies
	1.9 Capturing Metadata with or Without Provenance
	1.10 Summary
	References
2 Data Provenance and Accountability on the Web
	2.1 Data Longevity
		2.1.1 The Good
		2.1.2 The Bad
		2.1.3 The Ugly
	2.2 Data Provenance
		2.2.1 Rights Expression
		2.2.2 Provenance System Implementations
		2.2.3 Limitations of Provenance
	2.3 Data Accountability
		2.3.1 Definition of Accountable Systems
		2.3.2 Implementations of Accountable Systems
		2.3.3 Ensuring Accountability on the Web
	2.4 Future Directions
	2.5 Conclusion
	References
3 The Right (Provenance) Hammer for the Job: A Comparison of Data Provenance Instrumentation
	3.1 Introduction
	3.2 Case Study: Machine Learning Pipelines and Orange3
		3.2.1 Provenance Needs in Orange3
	3.3 Overview of Instrumentation Possibilities
		3.3.1 Human-Supplied Capture
		3.3.2 GUI-Based Capture
		3.3.3 Embedded-Script Capture
	3.4 Comparison of Instrumentation Approaches
		3.4.1 Provenance Collected
		3.4.2 Answering Provenance Queries
		3.4.3 The Cost of Provenance Instrumentation
	3.5 Conclusions
	References
4 Contextualized Knowledge Graphs in Communication Network and Cyber-Physical System Modeling
	4.1 Introduction
		4.1.1 Heterogeneity Issues of Representing Cyber-Knowledge
		4.1.2 Introduction to Knowledge Graphs in Cybersecurity Applications
	4.2 Utilizing Cybersecurity Ontologies in Knowledge Graphs
	4.3 Graph-Based Visualization
	4.4 Using Knowledge Graphs in Security System and Cyber-Physical System Modeling
	4.5 Task Automation in Cyberthreat Intelligence and Cyber-Situational Awareness Using Knowledge Graphs
	4.6 Summary
	References
5 ProvCaRe: A Large-Scale Semantic Provenance Resource for Scientific Reproducibility
	5.1 Introduction
		5.1.1 Related Work
		5.1.2 Overview of the ProvCaRe Resource for Semantic Provenance for Scientific Reproducibility
	5.2 Development of the ProvCaRe Ontology
		5.2.1 The ProvCaRe S3 Model
		5.2.2 ProvCaRe Ontology
	5.3 A Provenance-Focused Ontology-Driven Natural Language Processing Pipeline
		5.3.1 ProvCaRe NLP Pipeline
	5.4 ProvCaRe Knowledge Repository and Provenance-Based Ranking
		5.4.1 ProvCaRe Query Interface
		5.4.2 Provenance-Based Ranking
	5.5 Discussion
	5.6 Conclusion and Future Work
	References
6 Graph-Based Natural Language Processing for the Pharmaceutical Industry
	6.1 Introduction
	6.2 Application Area 1: Topic Identification
	6.3 Application Area 2: Patient Identification
	6.4 Application Area 3: Clinical Decision Support
	6.5 Application Area 4: Pharmacovigilance
		6.5.1 Graph-Based Natural Language Processing to Detect Adverse Events from ICSR
		6.5.2 Graph-Based Natural Language Processing to Detect Adverse Events from Electronic Health Records (EHRs)
		6.5.3 Graph-Based Natural Language Processing to Detect Adverse Events from Scientific Literature
		6.5.4 Graph-Based Natural Language Processing to Detect Adverse Events from Digital Content
	6.6 Knowledge Graph Development and Learning
	6.7 Retrofitting Distributional Embeddings with Relations from Knowledge Graphs
	6.8 Graph-Based Natural Language Methods from Outside the Pharmaceutical Industry
	6.9 Conclusion
	References