Reproduce “NSIDDx: A Design Framework for Neuro-Symbolic, Practitioner-First Differential Diagnosis in Low-Resource Settings” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Aarav Singh
NSIDDx pairs an LLM with a symbolic layer to make symptom evidence and diagnostic candidates inspectable and editable by practitioners. In a researcher-led sarcoidosis demonstration, correcting inferred symptom absences and adding missed splenomegaly ranked sarcoidosis first. On 500 CUP cases, DDx@5 semantic accuracy was 75.2%, versus 6.6% exact accuracy. This gap depends on the matching rule; the pipeline was evaluated on case reports, not clinician decisions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.