Program Festivalday 1

Panel

Transparency & intelligence? Advancing artificial intelligence with Open Science

This panel addresses a vital question in our current times: What challenges and opportunities arise at the intersection of Open Science and Artificial Intelligence? How can the principles of Open Science be applied to the field of artificial intelligence? To address these and related questions, a group of speakers working at the intersection of science, technology and society will debate challenges and practices to be considered when integrating Open Science into AI technologies.

Speaker: Clara Bersch, Psychology and Computer Science, Max-Planck-Institut for Human Development; Prof. Dr. Kristian Kersting, Artificial Intelligence and Machine Learning, TU Darmstadt; Dr. Katja Mayer, Science and Technology Studies, University of Vienna; PD Dr. Andreas Sudmann, Media Studies, Rhenish Friedrich Wilhelm University of Bonn

Moderation: Carolina Natel de Moura, Digital Research Academy


Workshops

What is Open Science?

“Open science: just science done right?” (Tennant, 2019). But what exactly does that mean? This workshop aims to demystify Open Science. This term is increasingly used in academic circles, yet it often lacks a clear definition. This workshop is suitable for those new to the scientific field or to Open Science itself.
This session aims to provide a clear explanation of what Open Science entails and discuss its core principles. This workshop will feature discussions and activities to help you develop practical skills and strategies that you can apply in your daily research activities to enhance transparency and collaboration, thus adopting a standard practice of Open Science.

This is an opportunity to gain a thorough understanding of Open Science, connect with fellow researchers, and discuss common challenges and solutions. Join us to learn how you can apply Open Science principles to improve your work and contribute to a more open scientific community.

Workshop leader: Merle-Marie Pittelkow, postdoctoral researcher QUEST center – BIH at Charité Berlin


Open Science meets Mastodon: how the Fediverse connects science and how you can be a part of it

In recent years, many researchers and scientific organizations have left Twitter (now "X") for good reasons and migrated to the Fediverse, a community-oriented federated network that invites a more self-determined, human interaction. The Fediverse operates on a donation-based, non-commercial model without advertisements or the sale of user data. The largest and most well-known project in this space is currently Mastodon.

We will begin by exploring the Fediverse together. Participants will create a Mastadon account (unless they already have one), adjust initial settings, try out different methods of interaction within the Fediverse, and explore various hubs of the scientific community.

In the second part of the workshop, we will delve into advanced topics such as alternative interfaces, discovery, trend monitoring, account migration, post automation, cross-posting, groups, bots, bridges, instance hosting, and other practical aspects that are useful for science communication. Participants are welcome to bring their own needs, questions, concerns, or experiences with Mastodon / the Fediverse into the discussion.

The workshop agenda and topics can be viewed in advance on Cryptpad (https://cryptpad.fr/pad/#/2/pad/edit/bQT-TCfMv3ILK8tTcf23oIOj/), and participants are encouraged to add their suggestions or comments. Please bring a device (ideally a laptop) with internet access to the workshop.

Workshop leader: Nicolas Bach, student Information Science, University of Applied Science Darmstadt


Statistical inference and data visualization for the humanities – live coding and discussion

The world of data is opening up day by day, facilitating the unprecedented access researchers have to big data. This enhanced access can help researchers infer important, interesting and groundbreaking knowledge from data that will in turn serve further theoretical and methodological development. At the heart of this journey lies data visualization – a critical tool not just for initial statistical analysis, but also for data cleansing and preparation for more advanced processes like machine learning and deep learning.

Yet, without a foundational understanding of statistics, researchers risk missing out on the full power of data visualization. This is firstly because statistical knowledge determines which tools and methods should be used to visualize specific datasets and address particular research questions. Secondly, scholars equipped with basic statistical knowledge can better interpret graphs and charts representing certain datasets.

Dispelling the myth that statistics and coding are insurmountable barriers for scholars in the humanities, this workshop invites coding novices to join us in a collective live coding experience. Even without expertise in Python or mathematics, participants can benefit from this workshop and take home creative ideas to embark on their journey of coding, tackling big data problems in the humanities.

The live coding segment will be conducted via Google Colab, so participants only need an internet browser and no pre-installations are necessary. We will utilize the Python library Seaborn to create appealing visualizations from a dataset prepared by the workshop leaders. Participants will grasp fundamental concepts of statistical inference and gain practical skills in implementing this knowledge to either visualize data themselves or interpret data visualizations created by others. In the latter part of the workshop, we will partake in group discussions to explore how these newly acquired skills can elevate the art of visual storytelling with data, making the interaction with large open datasets and the insights gained both accessible and engaging.

This workshop targets experts from all areas of humanities working with big data, preferably with a basic knowledge of Python. If you have never written code in Python but you are eager to begin learning it, you are also most welcome at the workshop.

Workshop leader: Golnaz Sarkar Farshi, research associate Philipps University Marburg

Workshop team: Julia Tolksdorf, research associate University of Applied Science Mainz; Robert Zwick, research associate University of Applied Science Mainz


Systemic changes in sight?! Exploring the complexity of crediting in Open Science

Reputation is crucial for advancement in the academic system. The evaluation of researchers depends less on the quality of their research and more on the quantity of their publications and the reputation of the journal in which they were published. Recently, the discourse on scientific assessment has extended beyond publications to acknowledge the multidimensionality of scholarly work and output. This pertains not only to the significance of teaching but also largely to reviewer activities.

The Student Network for Open Science –  a student-run open science journal, education platform, think tank, and research hub – advocates for a comprehensive transformation of current research practices to embrace Open Science as the new standard in science and knowledge dissemination. As early career researchers, we see the need to adjust the current crediting system. Only by incentivizing good scientific practices can the system be transformed from a closed system to an Open Science system. However, change is not always easy. 

During our workshop, we want to delve into the complexity of crediting by asking questions such as: 
(i) Is recognition necessarily bound to the individual researcher?
And (ii) how can we make the system work for a practical shift towards Open Science principles?

The workshop consists of three major parts:
(i) Stakeholder analysis in the light of their role in the current scientific practice [e.g., established researchers, ECRs, funders, universities, publishers];
(ii) System mapping: connecting the stakeholder positions;
(iii) Identification of target points for systemic change.

Workshop leaders: Felix Hambitzer; Amelya Keles; Anne Neumann, Student Network for Open Science


Hands-on semantic data management with LinkAhead: increased data findability and reusability

Data management involves the storage, browsing, retrieval, and analysis of datasets and their connections and circumstances. Good data management makes valuable data reusable for current and future users. It also makes data findable ("Where is the training data for sensor X of setup Y again?") and adds real utility to data, because data can be embedded into context ("Which experimental settings were used for obtaining the data for project P, and how many failures were there?"). Thus, good data management supports preparing FAIR open data that is useful to everyone with whom it is shared.
The open-source toolkit LinkAhead (formerly known as CaosDB) represents the practical implementation of an agile data management approach designed to handle all these tasks and much more. The structure of data can be modified later without losing old information and without the need to migrate existing data. This encourages agile implementation of data management workflows instead of delaying until the perfect master plan is ready.

And LinkAhead comes with a powerful Python client, so access is as easy as a few lines of code. This session consists of a short live demonstration of the LinkAhead Python client, and participants are encouraged to follow along on their own computers. To do so, they can install LinkAhead's Python library and additional tools with "pip install linkahead caosadvancedtools" and make sure that they can load the library in Python with "import linkahead."
A Jupyter notebook will be made available online before the session.

Workshop leader: Dr. Daniel Hornung, IndiScale GmbH

Workshop team: Dr. Alexander Schlemmer; Dr. Florian Spreckelsen; Dr. Henrik tom Wörden, IndiScale GmbH


How much is the fish? - OpenCost for cost transparency and sustainable and fair scientific publishing

The publication of articles, books, data, software, or any other scientific output is not free of charge. Often, researchers must use their own funds to pay invoices from service providers. As a result, these invoices can easily get lost in the accounting systems of research organizations, even though they consume a considerable portion of the research budget.

To counter this risk, costs must be stored as metadata alongside the research findings and made publicly available in a standardized and machine-readable form. To this end, openCost has developed an exchange format and proposes the well-established OAI-PMH interface to expose it from a repository, CRIS, or any other suitably equipped systems.

As the name suggests, the openCost format is openly available and can model any cost type in any publishing model, thus facilitating cost transparency. Besides external costs (e.g., APCs), the current schema also allows for the modeling of internal costs, such as those for storage, long-term archiving, staff, or indexing services.

The workshop will begin with a brief presentation of the openCost schema and its constituent parts. However, the majority of the time is reserved for discussion with the participants. We want to hear about your use cases and requirements for cost mappings in the context of Open Science with the aim of integrating them into the schema.

Workshop leaders: Dr. Gernot Deinzer, OpenCost; Dr. Alexander Wagner, OpenCost


Avoiding a new digital divide: Wie können kleine und mittlere Hochschulen im Forschungsdatenmanagement mithalten? (Deutsch)

Professionelles Forschungsdatenmanagement (FDM) hat zunehmend die Aufmerksamkeit von Forscher:innen, Förderorganisationen, Forschungsinfrastrukturanbietern (wie Bibliotheken oder Rechenzentren an Universitäten), Forschungsservice Institutionen und politischen Entscheidungsträgern auf sich gezogen. Trotz gemeinsamer Bemühungen, das FDM zu professionalisieren und institutionalisierte FDM-Dienstleistungen zu etablieren, gibt es noch offene Fragen, die geklärt werden müssen, um eine neue „digital divide“ zu vermeiden (siehe Lythreatis et al., 2022, für einen aktuellen Literaturüberblick). Wir möchten insbesondere die folgenden Fragen im Workshop diskutieren:
Was können wir tun, um das Entstehen einer neuen digitalen Kluft zu vermeiden? Sollten beispielsweise Universitäten, die bereits Archivlösungen besitzen, ihre Archive für externe Forscher öffnen, oder sollte es eine Finanzierung für Archiv-Infrastrukturen an anderen Universitäten geben? Wie können technische Lösungen gestaltet werden?

Wer kann sich langfristig und strategisch mit seinen FDM-Aktivitäten/Dienstleistungen beschäftigen, und wie steht es um Data Stewards in befristeten Verträgen, die sich oft auf kurzfristige Ziele konzentrieren?

Um diese Fragen im Workshop zu diskutieren, laden wir ein breites Spektrum von Teilnehmenden ein. Data Stewards, Forschungsinfrastruktur-Anbieter oder Praktiker:innen, die am FDM interessiert sind, sind ebenso eingeladen wie Forscher:innen, die ihre Anforderungen an das FDM teilen möchten. Wir freuen uns auch über internationale Teilnehmende, die ihre Erfahrungen beim Aufbau von FDM-Diensten in anderen Ländern teilen möchten.

Workshopleitender: Dr. Johannes Putzke, Data Steward Informatik, Hochschule Trier

Workshopteam: Lucas Krajewski, wissenschaftlicher Mitarbeiter Hochschule Koblenz; Jörg Panzer, wissenschaftlicher Mitarbeiter Hochschule Kaiserslautern; Manuela Richter, Hochschulleitung und Verwaltung Hochschule Mainz; Dr. Thomas Schimmer, Hochschulleitung und Verwaltung Hochschule Mainz


Geodata in Linked Open Data and Wikidata in Practice

Linked Open Data (LOD) is an essential component of modern RDM initiatives for the FAIRification and comprehensible modelling of research data in the spirit of Open Science. Especially in the NFDI – and its various consortia – ontologies and the resulting LOD play a decisive role in linking heterogeneous and interdisciplinary research data. The W3C standard used for this purpose – the Resource Description Framework (RDF) – already offers a diverse toolkit to enable data to interact with each other in an interoperable manner. Geodata in particular is an entity that brings different data together.
In addition to its established standards from the OGC service family (WMS, WFS, OGC API features, ...), the Open Geospatial Consortium (OGC) has also implemented a standard for geodata in RDF: GeoSPARQL.

However, the creation, publication and processing of (geo-)LOD in daily work, e.g., GIS, is not trivial and requires some expertise. In addition, community hubs such as Wikidata (Wikimedia) offer great opportunities for the participation of citizen scientists and the integration of data from Citizen Science. The SPARQL Unicorn was developed as a QGIS plugin [https://t1p.de/sparqlunicorn-qgis], which is capable of importing RDF data into the local GIS via SPARQL queries, to integrate this into daily working practice. It also offers options for transforming geodata into linked data and for generating HTML documentation of ontologies and research data [https://t1p.de/sparqlunicorn-ontdoc].

This workshop provides basic knowledge of geodata as Linked Open Data, the SPARQL query language and Wikidata. In addition, FAIRification tools such as SPARQL Unicorn will be introduced to integrate geodata modelled as LOD in daily work, e.g. with GIS. The workshop relies on data from the humanities and cultural sciences, in particular from archaeology.

Workshop leaders: Timo Homburg, research associate, University of Applied Sciences Mainz; Florian Thiery, Research Software Engineer, Leiza


Quarto im Forschungsdatenlebenszyklus: Transparente Dokumentation von der Datenerhebung bis zur Publikation (Deutsch)

Quarto ist ein innovatives Open-Source Werkzeug, das die Publikation wissenschaftlicher Dokumente mit reproduzierbaren Analysen vereinfacht. Durch die Möglichkeit dynamischen Output (Methoden, Daten) und Texte in ein Dokument zu integrieren, wird die Nachvollziehbarkeit der Forschungsergebnisse gesteigert. Zusätzlich basiert Quarto vollständig auf Pandoc Markdown und unterstützt viele verschiedene Ausgabeformate wie zum Beispiel HTML, MS Word, MS PowerPoint, PDF, EPUB, sowie die Erstellung von Dashboards und Websites. Insgesamt bietet Quarto eine flexible Plattform, die es Forschenden ermöglicht, mit verschiedenen Programmiersprachen (R, Python etc.) zu arbeiten und komplexe wissenschaftliche Inhalte zu erstellen, zu formatieren und zu veröffentlichen.

Neben einer Vorstellung der Funktionalitäten in Quarto soll der Workshop auch Raum geben, um gemeinsam zu diskutieren, inwiefern Tools wie Quarto einen festen Bestandteil in der Vermittlung von Open Science Praktiken bilden sollten. Welche Rolle die Verwendung solcher Tools für die Reproduzierbarkeit in der Forschung momentan und in Zukunft spielen werden.
Wie Tools wie Quarto dazu beitragen können die Lücke zwischen Forschen und Veröffentlichen zu schließen.

Workshopleitender: Prof. Dr. Kai-Christian Bruhn, interdisziplinäre Anwendungen raumbezogener Mess- und Informationstechnik, Hochschule Mainz

Workshopteam: Lennart Linde, wissenschaftlicher Mitarbeiter Hochschule Mainz


Offenes Kulturerbe durch Citizen Science (Deutsch)

Was ist Kultur, was soll bewahrt werden – und wer trifft diese Entscheidungen? In vielen Haushalten werden familiäre Erinnerungsstücke wie ungeöffnete Schätze auf Dachböden aufbewahrt. Diese persönlichen und privaten Archive umfassen eine Vielzahl von Dokumenten wie Briefe, Urkunden, Alben, aber auch Fotografien und Videos, die von vergangenen Zeiten Zeugnis ablegen und die Kulturträgerschaft einzelner repräsentieren. Doch leider fehlt oft das Bewusstsein für diese individuelle Kulturträgerschaft, die Bedeutung solcher Schriftstücke und Ego-Dokumente und für deren interdisziplinäres Forschungspotential. Mit Citizen Science, d.h. mit bürgerwissenschaftlicher Beteiligung interessierter Personen, können diese privaten Archive und das enthaltene Kulturgut geöffnet und in einen open process gemeinsamer wissenschaftlicher Erschließung, Interpretation und Diskussion eingebunden werden.

Der Dialekt steht als immaterielles Kulturgut ebenfalls nicht im Fokus des öffentlichen Interesses, wenn es um die Bewahrung von kulturellem Erbe geht. Gleichzeitig geht die Zahl der kompetenten Dialektsprecher*innen deutschlandweit zurück. Für die Erschließung von dialektalen Sprachdaten in Kurrentschrift aus dem 19. Jahrhundert ist die Wissenschaft auf genau diese Dialektsprecher*innen aber dringend angewiesen. Nur mit ihrer Hilfe kann dieser große Schatz an dialektalen Sprachdaten gehoben werden und der Forschungsgemeinschaft zugänglich gemacht werden.

In diesem Workshop möchten wir mit den Teilnehmer*innen ein Bewusstsein für die Bedeutung des eigenen kulturellen Erbes näher bringen, Potential und Relevanz für Erforschung und Bewahrung von bislang marginalisierten Kulturgütern hervorheben sowie erprobte Tools (digitale Instrumente), vorhandene Perspektiven und Erfahrungen im Bereich von Citizen Science erörtern. Dabei werden Herausforderungen und Grenzen, aber auch Chancen eines offenen Kulturerbes im open process diskutiert.
Leitende/Team des Workshops kommen aus dem Citizen-Science-Projekt Gruß & Kuss und dem Liebesbriefarchiv, sowie aus dem Forschungszentrum Deutscher Sprachatlas in Marburg, der unter anderem die Wenkerbogen-App betreibt.

Durch diese Hintergründe kann auf umfangreiche Praxiserfahrungen hinsichtlich Citizen Science und interdisziplinäre Forschung im Bereich des Kulturerbe-Sektors zurückgegriffen sowie wertvolle Transfermöglichkeiten in Wissenschaft und Gesellschaft benannt werden.

Workshopleitende: Nadine Dietz, wissenschaftliche Mitarbeiterin Technische Universität Darmstadt; Dr. Lisa Dücker, wissenschaftliche Mitarbeiterin Phillips Universität Marburg; Lena Dunkelmann, Projektmitarbeiterin Gruß und Kuss, Universität Koblenz; Dr. Robert Engsterhold, wissenschaftlicher Mitarbeiter Phillips Universität Marburg

Workshopteam: Prof. Dr. Andrea Rapp, Germanistik - Computerphilologie und Mediävistik, TU Darmstadt; Prof. Dr. Eva Lia Wyss, Sprachwissenschaft und Sprachdidaktik, Universität Koblenz-Landau