Conceptual Structures for the Digital Humanities
by Tom Hanika, Sergei Obiedkov, and Robert Jäschke, Monday, September 11, 09:30-11:00
The tutorial will explore the question, whether formalized conceptual structures are vital to model and analyse data for the digital humanities. Three examples – ranging from modelling character structure in dramas using formal concept analysis to topic trajectories for textual analysis – will pave the ground for a subsequent discussion. Together we want to explore, whether and how conceptual structures, methods and tools can support research in the (digital) humanities.
Needs beyond Chat-GPT: Teaching the Concepts of Knowledge Representation by Filling a Void
by Jan Krämer and Lilian Löwenau, Monday, September 11, 11:30-13:00
In this world filled with hype and breathless adoration for the abilities of large language models (LLMs), it is a challenge to keep in mind the actual limitations of these abilities – even if those limitations are currently not yet fully understood. We are interested in how this interest, and the limitations, can be used as a teachable moment and source for educational resources regarding more structural and semantic approaches.
After an introductory presentation on the current state of the discussion regarding LLMs, we invite all participants to share their ideas as well as examples from current coursework. We aim to use this discussion to create a concise overview of the topic to be shared among the participants.
The R package fca-R
by Domingo López-Rodríguez and Ángel Mora Bonilla, Monday, September 11, 14:30-16:00
The fcaR package is a comprehensive library developed in the R programming language, designed to facilitate the application of Formal Concept Analysis (FCA) techniques. It encompasses the essential structures of FCA, such as formal contexts, concept lattices, and implication systems, along with relevant algorithms for their manipulation. Notably, fcaR goes beyond traditional FCA by incorporating fuzzy datasets and employing Simplification Logic for automated reasoning.
As the first R-based implementation of FCA techniques, the fcaR package has gained significant traction, with over 21,000 downloads from the official repository. It has been extensively utilized in scientific research, appearing in several works focusing on recommender system creation and modeling the knowledge space of mathematics courses to aid educators in understanding content dependencies. Additionally, fcaR has been employed for theoretical investigations, allowing researchers to swiftly test algorithms and present novel extensions, such as the simplification of implication systems involving positive and negative attributes.
Given R’s status as a popular data-oriented programming language alongside Python, the fcaR package offers a compelling opportunity to introduce the FCA framework to diverse communities, including data scientists, engineers, psychologists, and more. By leveraging the capabilities of fcaR, these groups can effectively address knowledge extraction problems and benefit from the FCA paradigm.
In this tutorial, participants will gain hands-on experience in harnessing the capabilities of the fcaR package for Formal Concept Analysis. The tutorial will commence with an overview of the fundamental structures and algorithms employed in FCA, providing a solid foundation for understanding the package’s functionalities. Participants will then delve into practical examples, exploring how fcaR can be used for tasks such as formal context creation, concept lattice generation, and implication system manipulation. Through interactive demonstrations, attendees will learn how to apply fcaR to real-world scenarios.