Track Machine Learning in Your Research Domain

Machine learning (ML) is becoming increasingly important across research disciplines, driven by advances in AI, especially large language models (LLMs). As part of the NFDI consortia BERD and NFDI4DataScience (NFDI4DS), we develop methods to automatically extract ML-related information from research publications. Our pipeline, based on GSAP-NER, identifies ML models, datasets, tasks, and methods in scientific texts. A user-facing service enables researchers to upload PDFs and receive highlighted terms and aggregated statistics. This structured access improves research data management and supports reproducibility. Future work includes linking extracted entities to external knowledge graphs and enabling cross-domain comparisons of ML usage.