Open Password – Monday June 13, 2022
#1085
Open Science – Knowledge Graphs – Linked Open Data – Ontologies – ZPID – Tina Trillitzsch – Linked Open Data Cloud – ZPID Knowledge Graphs – PSYNDEX – PubPsych – PsychArchives – Open Knowledge Graph – SPARQL – Semantic – Entities – Relationships – Properties – Semantic Network – Semantic Search engine – Google’s Knwledge Graph – Amazon’s Alexa – Wikidata – Psychology – Resource Description and Access – Semantic Modeling -PsychPorta – PsychAuthors
KOALA – Community Open Access Financing – Media Studies – Social Sciences – TIB – KIM – Article Processing Charges – Marco Tullney – Sören Auer – KOALA consortia – Minimum standards – Plan S – Price tiering – BMBF
Open science
Knowledge graphs, linked open data and ontologies for the psychological data sets
of the Leibniz Institute for Psychology (ZPID)
By Tina Trillitzsch
Tina Trillitzsch
_____________________________________________________________________
Summary
_____________________________________________________
It is becoming increasingly important to make scientific data available as open knowledge graphs on the Semantic Web in order to comply with open science principles and to make it easier to find, classify and reproduce studies. Open knowledge graphs make the relevant data from studies and publications freely available to anyone interested, understandable for semantic search engines and enable linking to unique entities for responsible people and institutions.
This article explains the core terms on the topic and explains how the Leibniz Institute for Psychology (ZPID) will offer its databases from the field of psychology in this form. After a brief overview of knowledge graphs and their possibilities for semantically representing information, an introduction to Linked Open Data, the Linked Open Data Cloud (LOD Cloud) and the role of a future ZPID knowledge graph within it follows. Finally, the fundamental role of ontologies for knowledge graphs is discussed and the ZPID’s own ontology is discussed. This is currently under construction. Goals, the current status and plans for further development are discussed.
_____________________________________________________
The ZPID and its planned knowledge graph to provide psychological data
_____________________________________________________
The ZPID (Leibniz Institute for Psychology) is an infrastructure facility for psychology in German-speaking countries. It maintains the literature database PSYNDEX for psychological research literature, the search portal PubPsych for free searching of PSYNDEX and the repository PsychArchives for the long-term archiving of psychology-relevant digital research objects such as articles, research data, presentation slides, etc. In the future, ZPID will make these previously largely separate data sets available as open knowledge graphs and so on in particular, increase the discoverability of psychological information, studies and research data and make it easier to use them. The graph will enrich data with semantic meaning and network the isolated “data silos” internally, but also create rich external links to other free data sets and authority files on the Internet. The graph becomes interoperable through the use of open standards and formats such as RDF, SKOS and OWL as well as through the support of the free knowledge graph interface standard SPARQL. In the future, users and other services will be able to query data from different sources together, combine them and use them flexibly in innovative applications. For this purpose, ZPID is currently developing an ontology that will serve as a “blueprint” for the knowledge graph.
_____________________________________________________
What is a knowledge graph?
_____________________________________________________
A knowledge graph is a method with which data is stored in a structured form and its meaning ( semantics ) is explicitly recorded. In addition, the data is stored and linked together like a flexible spider web instead of being strung together as rigid tables in a database as is usually the case.
Structured means that the information is divided into individual, predefined elements. Co-capture of semantics , which means: For each of these captured information elements, it is described exactly what type of information it is (e.g. “person”, “has first name”). This is best done in multiple languages, with explanations and synonyms.
We need this explicit meaning enrichment of the data so that computers can really «understand» it and search and link it in a meaningful way – because they cannot derive the meaning of things from the context like humans do. You first have to explicitly tell them all the meanings. An example of an unstructured text without explicit semantic information is the following statement:
«[…] Michael Bosnjak […] took over the management of the ZPID.» (Excerpt from Wikipedia article on ZPID)
People know from experience what is meant here. Computers don’t understand many things in this excerpt: What is a «ZPID»? What a “Michael”, what a “line”? Does this mean a power line or the leadership position of an organization? How are these things related to each other ? Therefore we have to use the meaning “to say”.
_____________________________________________________
How semantics are represented in knowledge graphs
_____________________________________________________
“Things” are stored as so-called entities : as individuals of a certain type. These entities are linked to each other through named relationships . Simple properties of things (text values such as years or names) are also linked to their entities via relationships.
This turns the unstructured text «[…] Michael Bosnjak […] took over the management of ZPID» into the following semantic structure:
- “ ZPID ” is an entity of type “ Organization” .
- This ZPID organization entity has a relationship with the entity «Michael Bosnjak» of type Person , namely a relationship of type «has director of an organization» (In addition, we keep an English name of the same relationship, as well as a more detailed explanation and possible synonyms).
- The ZPID entity also has a relationship to the date value «1972», and the type of this relationship is «has year founded»
This information comes together to form a structured network enriched with meaning, a “semantic network” , with entities as nodes. The “threads” form the relationships to other entities or to values. In order to save something new to the «ZPID», you «stretch» another (named) thread to another node or a new data value.
To query the data stored in this way, we use a semantic search engine that knows all the links and their possible names and explanations. She “shimmies” her way through the network from node (entity) to node, always along its threads, until she finds what she is looking for and presents it to us.
Below we ask two questions to show how a semantic search engine answers them by navigating an extended version of our network:
- Who is Director of the Leibniz Institute for Psychology? – Search all entity names, find the appropriate one using the “has main name” relationship, from there to the ZPID node itself, then to the person M. Bosnjak using the director relationship.
- What does M. Bosnjak write about? – Find a person entity with the given name, see if it is linked to publication entities with the “is author of” relationship; then from there check whether the publications have relationships of the type «has topic» and list which keywords are at the end of them.
Tina Trillitzsch is a computer scientist in the information & research department at ZPID (Leibniz Institute for Psychology). She has worked in the areas of conceptualizing information architectures and front-end web development and has many years of experience with literature databases in the field of psychology. She is currently developing an ontology for the semantic indexing of ZPID’s databases, including the literature database PSYNDEX. She is also responsible for the development and maintenance of the help and information website psyndex.de responsible.
Read the final episode: Knowledge graphs in the “wild” – plans and goals after the ontology is completed
Project KOALA
Seeking collaborative
open access funding
(TIB) Institutions can now participate in collaborative open access funding in media studies and social sciences
In the BMBF-funded project KOALA (“Building Consortial Open Access Solutions”), the starting signal for the first pledging phase is given with publications from the fields of media studies and social sciences. The aim of the two project partners, the TIB – Leibniz Information Center for Technology and Natural Sciences and the Communication, Information, Media Center (KIM) of the University of Konstanz, is to jointly establish sustainable, collaborative financing of open access journals and series of publications. which enables operation without costs for authors .
In the so-called pledging phase, supporters, libraries, research institutions, foundations, museums, associations, companies and individuals undertake to contribute to the costs of financing open access publications. When the specified number of supporters is reached, they make a contribution based on their size and the scope of the bundled publications. If there are not enough supporters, KOALA will not provide funding.
_____________________________________________________
How does the KOALA model work?
_____________________________________________________
Traditionally, scientific journals and series of publications are financed through (usually institutional) subscriptions or the sale of individual copies. This excludes many potential readers. Many open access journals, on the other hand, charge authors costs; they charge so-called Article Processing Charges (APCs).
“We want free access to scientific knowledge without financial hurdles. That’s why we strive for sustainable and fair open access models,” says Marco Tullney, KOALA project manager. Prof. Dr. Sören Auer, director of the TIB, explains: “The joint financing of open access publications through broad-based consortia follows on from the extensive and good experiences with library consortia. We are now aligning this instrument with open access.”
KOALA’s pilot portfolio includes the following two bundles:
KOALA Media Studies 2023
- Moving images: https://www.buechner-verlag.de/reihe/bewegtbilder/
- Yearbook of Moving Image Studies: https://www.buechner-verlag.de/reihe/yomis/
KOALA Social Sciences 2023
- Forum Qualitative Social Research: https://www.qualitative-research.net/index.php/fqs
- GENDER: https://www.gender-zeitschrift.de/start
- Open Gender Journal: https://opengenderjournal.de/
- sub\urban: https://zeitschrift-suburban.de/sys/index.php/suburban/about
_____________________________________________________
What are suitable magazines and series for KOALA?
_____________________________________________________
KOALA consortia finance journals and series that provide immediate access to full texts and do not charge publication fees. Candidates include those that are already published open access as well as those that are currently only accessible to subscribers and want to change their model. Magazines or series of publications that currently still collect APCs but want to abandon this practice are also suitable.
In addition to the willingness to publish open access, KOALA requires the fulfillment of a number of minimum standards. These are aligned with Plan S, a strategy to promote open access to publicly funded scientific knowledge, and other relevant initiatives. The editorial teams and publishers involved in KOALA receive extensive support in implementing these standards.
_____________________________________________________
How are the prices determined?
_____________________________________________________
The prices for the individual bundles are determined based on the financial needs of the magazines and series that they report to KOALA for the sustainable operation of their publication. For each bundle there is a price tier (“tiering”) that participating supporters can assign to according to their size. The magazines and series will then receive funding from the KOALA consortium for three years.
_____________________________________________________
How can I or my institution participate in the financing?
_____________________________________________________
The majority of the acquisitions departments of academic libraries in Germany are currently receiving information about participation in the KOALA consortium by email. Other interested parties should contact the project team by email at koala@tib.eu or take part in one of the information events .
_____________________________________________________
Will other magazines be included?
_____________________________________________________
The compilation of magazines and publication series for the current pledging phase is final. Preparations for the next round of financing from 2024 are already underway. Interested journals or series are welcome to get in touch; all disciplines are welcome.
The KOALA project is funded by the Federal Ministry of Education and Research (BMBF) for a period of two years.
OpenPassword
Forum and news
for the information industry
in German-speaking countries
New editions of Open Password appear three times a week.
If you would like to subscribe to the email service free of charge, please register at www.password-online.de.
The current edition of Open Password can be accessed immediately after it appears on the web. www.password-online.de/archiv. This also applies to all previously published editions.
International Cooperation Partner:
Outsell (London)
Business Industry Information Association/BIIA (Hong Kong)
Open Password Archive – Publications
OPEN PASSWORD ARCHIVE
DATA JOURNALISM
Handelsblatt’s Digital Reach