Medical reasoning
Differential diagnosis
by formal evaluation
A patient describes symptoms. FR-OS checks 30,981 diagnoses against formally proven rules and returns a ranked differential with exclusion certificates showing why each diagnosis was ruled in or out. Every ruling is inspectable.
See it in action
FR-OS DDx Engine
A patient describes crushing chest pain, sweating, and nausea. FR-OS identifies the clinical findings, weighs all candidate diagnoses, and returns a ranked differential. No physician input required.
By the numbers
Verified medical knowledge
Each with Coq-checked rules derived from medical knowledge bases.
100% hit rate on curated vignettes. 78.4% on out-of-sample USMLE / MedQA questions, engine-only.
The engine is logically precise by construction. Every exclusion is a proof the clinician can inspect.
How it differs
Every ruling out comes with a certificate
AI-based diagnosis
Neural models produce ranked lists with probability scores. "Myocardial infarction: 73% likely." There is no explanation of why, no record of what was considered, and no guarantee the model won't produce a different answer tomorrow.
FR-OS diagnosis
The engine tests each diagnosis against the patient's findings using formally verified rules. Diagnoses are ranked by how much evidence supports them. Each exclusion comes with a certificate showing exactly which findings ruled it out. The same inputs always yield the same answer, with a full trace you can inspect.
Disambiguation in medicine
Words matter in clinical text
Medical language is full of ambiguity. "Cold" can mean temperature or illness; "discharge" can mean release from hospital or bodily fluid; "positive" can mean good news or a concerning test result.
FR-OS resolves these ambiguities before clinical reasoning begins. The same engine that achieves 94.5% on standard NLP benchmarks processes clinical text, ensuring that downstream diagnosis operates on resolved meanings rather than surface words. Every sense assignment and every diagnosis exclusion is produced by the deterministic engine; no embedding model or LLM sits in the decision path. This eliminates an entire class of errors where ambiguous terms match the wrong diagnoses.