Biomedical Expert Finding Benchmark: Health Claims Matched to Researcher Publication Profiles

This dataset provides the first large-scale benchmark for identifying biomedical experts qualified to verify health claims. It contains 93,404 health claims paired with 153,147 biomedical researchers, linked through their PubMed publication histories. The dataset addresses the critical challenge of connecting informal health claims from news media with formal biomedical literaturea cross-genre retrieval task essential for combating health misinformation. Each claim is associated with relevant experts based on their documented research expertise, creating gold-standard (15,119 claims) and silver-standard (78,285 claims) subsets with different construction methodologies. Expert profiles include comprehensive publication records from PubMed (2018-2024), with titles and abstracts representing their research domains. The dataset captures the semantic heterogeneity between how health information appears in public discourse versus scientific literature, making it valuable for developing automated expert identification systems. Applications include: Benchmarking expert matching and cross-genre retrieval systems Training models for health misinformation detection Analyzing expertise patterns in biomedical research Supporting fact-checking workflows in clinical and public health contexts Studying health communication between scientific and public domains