Folge verpasst? Kein Problem! In unserem Veranstaltungsarchiv finden sie sämtliche Ausgaben unserer Expert*innenreihe zum Nachhören. In den Informationen zur Veranstaltung finden Sie den Link zu den Aufzeichnungen auf Youtube. Zusätzlich haben Sie die Möglichkeit, die Powerpoint-Folien herunterzuladen.

Staffel 6: Knowledge technologies for the Social Science: Access to Social Science Data and Services

Die sechste "Meet the Experts"-Staffel präsentiert eine Reihe von Vorträgen aus der Abteilung "Knowledge technologies for the Social Sciences", die Forschungsdateninfrastrukturen und Technologien für die sozialwissenschaftliche Community bereitstellt, um den Zugang zu Forschungsdaten und Informationen mit Hilfe von künstlicher Intelligenz zu verbessern.

Die Satffel beginnt mit der Untersuchung des Informationssuchverhaltens und der Datenbedürfnisse von Sozialwissenschaftler*innen, die die Grundlage für geeignete Suchinfrastrukturen und Wissensgraphen bilden, um sozialwissenschaftliche Daten und Informationen auffindbar, zugänglich und wiederverwendbar zu machen. Darauf aufbauend wird gezeigt, welche Möglichkeiten und Herausforderungen sich aus der Übernahme technologischer Fortschritte in den Sozialwissenschaften ergeben, wie z. B. große Sprachmodelle und neuartige Datenquellen aus dem Web.

Abschließend wird die automatische Extraktion wissenschaftlicher Informationen aus wissenschaftlichen Texten vorgestellt, die dazu beiträgt, die Suche und das Verständnis von Forschungsinformationen und deren Abhängigkeiten zu verbessern.

Registration (via Zoom)   |  

Slides (3,61 MB) |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

GESIS conducts user research in order to provide user-friendly and high-quality access to information and data relevant to empirical social research. Part of these activities is to understand the information-seeking behavior of social scientists. In this talk, Dagmar Kern focuses on how information-seeking behavior is defined, why it is crucial to understand it and what methods exist to explore it. Furthermore, she will present what we have learned so far and how this is incorporated into a digital service you may know and use.

Presenters:

Dr. Dagmar Kern

Registration (via Zoom)   |  

Slides (1,11 MB) |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

Finding suitable datasets is a difficult task. But why? In this talk, we will look at national and international efforts to increase findability, why they are important and what everyone can do to make their data better findable. This will include concrete advice to game the system and increase your data citation count.

Presenters:

Dr. Brigitte Mathiak

Registration (via Zoom)   |  

Slides (3,76 MB) |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

In the social sciences, research data and related information are often distributed on websites, search portals, data archives, and databases. In this talk, we present GESIS Search, which provides a central search entry point to the information space of empirical social science. Users can find national and international research data sets, publications, survey variables, questions from questionnaires, survey instruments, and tools. We will talk about how research information can be found in different categories, how information is linked and can be browsed, and what the future will bring for GESIS Search. This talk addresses researchers and interested parties looking for data and interested in specialized search environments.

Presenters:

Dr. Daniel Hienert

Registration (via Zoom)   |  

Slides (3,41 MB)  |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

Sharing research data and information has become a crucial demand for reusability, reproducibility, and visibility. The FAIR principles give recommendations on how to improve findability, accessibility, interoperability, and reusability. In this talk, we will present the basic concept of Knowledge Graphs and will show how they can support the sharing of research data and information in a FAIR way.

Presenters:

Dr. Benjamin Zapilko

Debanjali Biswas

Registration (via Zoom)   |  

Slides (2,37 MB)  |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

Large Language Models (LLMs) have revolutionized natural language processing and offer exciting opportunities for the social sciences. This talk explores the potential of LLMs in enhancing research methods for data analysis and enrichment that are ultimately used to study social science research questions and in GESIS services. In particular, we look at how LLMs-based techniques such as zero-shot and few-shot learning can be applied to produce training data for methods for opinion mining, survey item tagging, and variable extraction. Alongside the opportunities, we also discuss the challenges and risks associated with the utilization of LLMs, including robustness, model decay, hallucination and privacy, and ethical issues. Finally, in this presentation, we also discuss the possible applications of LLMs for established GESIS services such as GESIS search but also for services under development, e.g., novel data offers created from Web content and social media or the GESIS Methods Hub, a new platform for sharing computational methods and models for analysis of digital behavioral data.

Presenters:

Dr. Dimitar Dimitrov

Dr. Hajira Jabeen

Registration (via Zoom)   |  

Slides (1,91 MB)  |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

Preserving data from social media is crucial for many scientific disciplines. Publicly available social media archives facilitate research in the social sciences and provide corpora for training and testing a wide range of machine learning and natural language processing methods. To reduce the reliance on commercial gatekeepers, we decided in 2013 to create a large-scale longitudinal archive of tweets from X (then Twitter) for research purposes. We collected data from the then freely available random sample of 1% of all tweets from Twitter’s streaming API. In this talk, we will introduce TweetsKB - a knowledge base of tweets that has been enriched with named entities and sentiments. We also show how TweetsKB can be used to create topic specific sub-corpora, focusing on important societal events such as the COVID-19 pandemic. Understanding the COVID-19 discourse, its differences to the general Twitter discourse, and interdependencies with real-world events or (mis)information can foster valuable insights.

Presenters:

Sebastian Schellhammer

Registration (via Zoom)   |  

Slides  |   Presentation on YouTube   |   MTE Playlist

The disscussion will be held in English.

In this panel discussion, we will focus on one important aspect of data quality: representation. To this end, survey experts will discuss the nuts and bolts of representation and sampling, including the notion of good representation, ways to achieve or to measure it, challenges on the way towards good representation, probabilistic and non-probabilistic samples, and innovative ways to supplement traditional survey data. Stay tuned for more information!

Participants:

Stephanie Eckman

Barbara Felderer

Peter Lugtig

Henning Silber.

Moderation by Jessica Daikeler and Ruben Bach

Registration (via Zoom)   |  

Slides (2,32 MB)  |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

Scholarly Information Extraction' involves identifying resources, concepts, actors, and their relationships from scholarly documents and related data sources, such as software repositories. This forms the basis for many use cases, including automated literature reviews and advanced search applications. The talk presents the fundamental concepts and methodologies for scholarly information extraction, followed by a presentation of a scholarly information extraction project from start to finish.

Presenters:

Wolfgang Otto

Dr. Lu Gan

Dr. Saurav Karmakar

Dr. Philipp Mayr