Jennifer D'Souza

I work as an NLP research group lead in the SCINEXT project and serve as a senior postdoctoral researcher for the Open Research Knowledge Graph project within the Data Science and Digital Libraries group at TIB.

My research focuses on neuro-symbolic systems for information extraction and knowledge representation, with the goal of transforming scientific literature into AI-ready, structured knowledge in knowledge graphs. A core part of my work is scientific schema and ontology engineering: I develop methods that combine large language models with structured representation paradigms to support reliable knowledge capture, comparison, and reuse across scientific domains.

I also contribute to the intersection of LLMs and knowledge organization through benchmark development and community leadership, including organizing shared tasks on LLM-based ontology learning (LLMs4OL at ISWC 2024 & 2025) and automated subject indexing (LLMs4Subjects at SemEval/GermEval). Complementing this, my team develops open-source tools and evaluation frameworks (e.g., OntoAligner, OntoLearner, schema-miner, SciKGExtract) that enable reproducible benchmarking research of LLMs in structured information extraction.

I hold a PhD from the University of Texas at Dallas, where I worked on relation mining from text with a focus on temporal and spatial relations. I later worked as a postdoctoral researcher at University of California, Davis, applying NLP to software engineering as part of the Naturalness of Software initiative, and subsequently spent time in industry developing concept discovery solutions.

My (selected) research interests are:

• Neuro-symbolic NLP for scientific information extraction

• Knowledge representation for science: knowledge graphs, ontologies, and schema/template design 

• LLM-assisted knowledge organization (alignment, normalization, linking)

• Human-in-the-loop workflows for reliable knowledge capture

• Evaluation and benchmarking for structured extraction and synthesis

• AI-assisted scholarly discovery and comparison in the ORKG ecosystem (ORKG platform and ASK ORKG) 

You can find an overview of my publications and projects on this website, or via the TIB research information system and Google Scholar. 

Last but not least :), if you’re curious about the Open Research Knowledge Graph (ORKG)—where I’m part of the research and development team—you can start with the short videos linked below. ORKG’s mission is to build a free and open knowledge graph of FAIR-compliant scholarly contributions, so research results can be structured, compared, visualized, and reviewed.

If you’d like to explore the platform, visit: https://orkg.org/
And if you want a more “search-first” entry point, our newer LLM-powered scientific search engine (ASK ORKG) is here:
https://ask.orkg.org/