The semantic Web and Linked Data

Linking Ontologies on the Web

  • How to Link Ontologies:
    • (1) Through URLs
    • (2) Mapping Terms

Nesting metadata within webpages

RDF: The Resource Description Framework

  • Semantic triples are beliefs written as triples of the web:

    • < object property value> or < subject predicate object >

  • Example:

  • < Leonardo painted LaMonaLisa >

  • < Shakespeare wrote Hamlet>

  • < Bob isInterestedIn LaMonaLisa>

  • RDF Syntax(RDFS), a simple logic language, is used to write triple, and can include instances/individual frames

    • Owl can be used to write them too, but used for ontologies on the web (generic frames/only classes)

  • Can be used to make graphs of links on the web

  • Allows for reasoning (i.e., you can query wikidata)

Triples on the web

  • Triples can be written on the web using:
    • Knowledge representation language e.g., OWL
      • typically only for generic frames/classes
      • also allows for beliefs with more complicated relations than in triple
    • Simpler logic language e,g, RDFS
      • also allows for individual frames/instances

Ontologies vs Taxonomies vs Vocabularies

  • Vocabularies: a set of terms to describe the metadata
  • Taxonomies: a hierarchical organization of terms
  • Ontologies: formal taxonomies defined using logic
    • Mathematical logic used to define what the hierarchies mean
      • Subclasses imply containment of instances
      • Instances of a class have all the properties of the class
    • Logic constraints
      • Example: all humans have exactly one biological mother

Provenance

Defining Provenance

  • Each time the workflow is executed, the system records the provenance of the results: what workflow was used, what its components were, what the input data was, and what values were assigned to the parameters.
  • Provenance can be part of metadata.
  • provenance for a resource(e.g, data, document, etc.) is a record that describes entities and processes involved in producing and delivering or otherwise influencing that resource.

The Importance of Provenance

  • Provenance fosters trust
    • Provenance provides a critical foundation for assessing authenticity and allowing reproducibility

Representing Provenance

  • PROV is used to represent Provenance

Three Uses of Provenance

  • Provenance as Process(Computing Steps, Actions, etc)
  • Provenance as Resources themselves
  • Provenance as Attribution (to people, institutions, etc)

Publishing Data

Problem with current practice

  • Data is often not made available in publications. Lack of reproducibility.
  • Data made available through investigator’s URL, but URL does not resolve.

Best practices to make data accessible

Data Citation

  • Data repositories and journals often specify how to cite data
  • Unique persistent Identifiers:main types of Unique Identifiers include:
    • Uniform Resource Locator (URL): minimal effort to create. No guarantee of persistence. Should not be used in papers.
    • Persistent URL (PURL): the same PURL can be resolved to different web address over time. It is easy to create your own PURLs, just remember to update whenever you move the data.
    • Digital Object Identifier (DOI): DOI can only be issued by a DOI authority. Data repositories can issue DOIs for data.

Accessibility

  • Manual Accessibility
  • Machine Accessibility

General metadata & Domain Specific meta data

Publication in a shared repository

Choose a License for data

  • Recommended: CC-BY and CC0

Publishing Software

Software citation

  • As with data, use a persistent unique identifier (PURL or DOI)
  • How to cite the software:
    • With an in-text pointer as you would cite any other paper (recommended)
    • With an in-text pointer in the “Acknowledgments” section

Choosing an License for Software

  • Copyright: automatically applied to software when it is created to grant the creator exclusive rights as an intellectual property
  • Open source license: reduce constraints and enable software developers to make their source code available to public.

Ideal data science reports

  • Data: available in a public repository, including documented metadata, a clear license specifying conditions of use, and citable using a unique and persistent identifier.
  • Software: available in a public repository, with documentation, a license for reuse, and a unique and citable persistent identifier.
  • Provenance: documented for all results with a workflow sketch and with a provenance record.

Metadata standards:

  • Friend-of-a-Friend (FOAF): focus is describing people and their relationships.
  • OWL Time: a proper ontology with detailed constraints about classes and properties.
  • The Gene Ontology (GO): focus is genomic information.
    How do standards come about?
  • Standards organizations
  • Industry-led
  • A PhD thesis
  • A community

The World Wide Web (W3C) Consortium:

  • Focus on web standards, such as HTML, XML, OWL, RDF, etc.
  • Incubators analyze state of the art, and generate proposals that become the charter of working groups.
  • Working groups have a finite amount of time to accomplish their charter and generate a standard.
    Need for standards: since there are many uses of provenance:
  • Open information systems
  • Science applications
  • Business practices