abstract
- The exponential growth of scientific publications challenges researchers conducting literature reviews, particularly in transforming unstructured formats like PDFs into structured, machine-actionable knowledge. While automation and large language models (LLMs) offer potential, concerns about trustworthiness and the necessity for human oversight remain. To address this, we present two complementary Human-in-the-Loop tools: ExtracTable and SciMantify. ExtracTable enables researchers to extract and organize knowledge from scientific PDFs into semi-structured tabular formats using a reproducible, modular framework. SciMantify extends this workflow by supporting semantic annotation tasks and integration of curated data into the Open Research Knowledge Graph (ORKG). Together, these tools illustrate a scalable, FAIR-aligned approach that leverages human and machine intelligence to facilitate high-quality, semantically enriched knowledge representation for literature reviews.