My research is about ...
... considering in particular (1) human agency and oversight, (2) transparency, and (3) diversity, non-discrimination, and fairness using ...
... using traditional methods as well as deep learning architectures (e.g., CNN, BiLSTM, Transformer).
I'm especially interested in using graph neural networks (GNNs) applied on knowledge graphs for search and recommender systems.
... representing knowledge in such a way that it is "understandable" by both humans and machines.
I'm especially interested in creating knowledge graphs (KGs) and using them for machine learning tasks.
... using methods, such as LLMs, for extracting information (e.g., entities, facts, events, and arguments) from texts and linking this information to knowledge graphs.
I'm especially interested in the semantic annotation of publications and news articles with arguments and bias labels.
Since 2017, I have pursued research on scholarly data mining. Among other things, I have worked on context-aware citation recommendation. This describes the task of recommending suitable publications for a given text (e.g., sentence in a paper). I also developed systems that recommend papers, datasets, and neural networks.
Furthermore, I have published scholarly knowledge graphs and approaches to scientific impact quantification. For instance, I have studied how to develop an h-index for datasets.
I develop AI methods for supporting mediators (e.g., at the UN) during peacemaking mediation processes. More information can be found in this article.
A system that recommends data sets given textual research problem descriptions.
Published at Sci-K'21.
A system that recommends in-text citations with explanations.
Published at ECIR'19.
A system to explore papers and analyze the way in which they are cited. When given a publication’s title or author, it provides a novel search function that allows users to retrieve all contexts in which the publication is cited, including an indication about citation polarity.
Published at JCDL'19.
A system for exploring the temporal trends of scientific concepts. Scientific concepts were captured by extracting noun phrases and entities from all computer science papers of arXiv.org.
Published in SWJ'18.
Crunchbase is an online platform collecting information about startups and technology companies, including attributes and relations of companies, people, and investments. By means of linked-crunchbase.org, we bring Crunchbase to the Web of Data so that its data can be used in the machine-readable format RDF by anyone on the Web.
Published at BIR@ECIR'22.
RefBee is an online system that retrieves the metadata of all publications for a given author from the various bibliographic databases and indicates which publications are missing in which database.
Published in QSS'21.
A knowledge graph with metadata about data sets from all scientific fields and links to publications in which the datasets are mentioned.
https://doi.org/10.5281/zenodo.3885249, https://lov.linkeddata.es/dataset/lov/vocabs/nno
Published at KGSWC'20
A knowledge graph with metadata about neural networks. We consider neural networks published on GitHub and based on the Keras framework. Our knowledge graph can be used for making neural networks FAIR (i.e., findable, accessible, interoperable, and re-usable). The FAIRnets Search (https://km.aifb.kit.edu/services/fairnets/) enables users to explore and analyze neural networks based on our knowledge graph.
Published at JCDL'23, Scientometrics'20 & BIR'19
A dataset based on all arXiv.org publications. Apart from providing the papers’ plain text, in-text citations were annotated via global identifiers. Furthermore, citing and cited publications were linked to the Microsoft Academic Graph, providing access to rich metadata. Our data set consists of over one million documents and 29.2 million citation contexts.
Published at ISWC'19
A large data set with over eight billion RDF triples with information about scientific publications and related entities, such as authors, institutions, journals, and fields of study. The data set is based on the Microsoft Academic Graph. We also provide entity embeddings for all 210M represented scientific papers.
SPARQL endpoint available at http://km.aifb.kit.edu/services/crunchbase-sparql/
An RDF dataset of Crunchbase as of October 2015, containing information about 1.9M jobs, 1.3M websites, 568k organizations, 520k news articles, 430k people, 60k products, and 300k acquisitions.
Feel free to contact me in case you would like to know more about my current research activities or if you are interested in any form of collaboration.