Evidence-first generation
Answer generation only follows an accepted EMR/CPG evidence bundle, rather than raw top-K context.
Clinical RAG · EMR + CPG · audit trail
MedSwin grounds clinical answers in patient records and guideline evidence.
MedSwin combines patient-specific EMR retrieval, clinical practice guidelines, biomedical reranking, specialist critique, and provenance tracking so each answer can be traced back to the evidence that shaped it.
Live evidence sonar
Clinical channels under audit
MedSwinevidence lockHover a beacon to inspect an evidence channel.
0B student
0B teacher
0agents
0% Recall@10
System principles
MedSwin separates retrieval, evidence quality, safety critique, and answer synthesis.
The system tracks provenance, calibrated relevance, facet sufficiency, contradiction handling, agent reliability, benchmark performance, and release artifacts as distinct parts of the clinical evidence workflow.
Calibrated inclusion
Reranker logits are converted to calibrated probabilities for threshold-based policy checks.
Safety is a first-class facet
Contraindications, interactions, exclusions, and contradictions remain visible throughout selection.
Auditability over confidence
Each claim keeps source metadata, score context, evidence grade, and facet role.
Architecture animation
Trace a clinical query through the evidence chamber.
A clinician query is decomposed into patient context and guideline evidence, reranked with biomedical relevance signals, checked for sufficiency, and returned with a grounded answer plus audit trail.
Pipeline playback
Clinician query enters the chamber.
Medical specialist
Training, distillation, and model merging are displayed as separate mechanisms.
The specialist model is built from supervised biomedical instruction data, teacher-student distillation, and training-free merge operators that control destructive interference between updates.
SFT mixture
Biomedical supervision sources
Hybrid SFT + KD
Teacher-student transfer
27BTeacher
hard labelstop-k soft logitsuncertainty transfer
7BStudent
Lt = α CE + (1 − α)τ² KL(pT(·|τ) ∥ pS(·))
Interference-aware composition
Merge operators reduce destructive update conflict.
Retrieval and sufficiency
Evidence is selected by clinical utility, not by raw top-K truncation.
Each retrieval stage uses a responsive diagram, and the sufficiency simulator animates toward threshold when more evidence is retrieved.
Facet sufficiency simulator
Build an evidence bundle
Critical facets are below acceptance threshold.
Multi-agent coordination
The MAC layer behaves like a specialist dive team.
Agents explore different hypotheses, return claim-level ledgers, and are aggregated through reliability-weighted evidence selection.
Claim-level ledger
Every claim keeps its source role.
| Facet | Role | Polarity | Grade | Trace |
|---|---|---|---|---|
| Guideline concordance | Recommendation | supports | CPG | doc · version · section |
| Patient applicability | Lab / comorbidity | qualifies | EMR | encounter · timestamp |
| Safety risk | Contraindication | conflicts | Safety | severity · source |
| Uncertainty | Contradiction pair | preserved | Mixed | adjudication status |
Contradictions are not averaged away.High-grade conflicts are preserved until the critic or final synthesiser explicitly adjudicates them.
Evaluation dashboard
QA, reranking, and audit metrics are inspected separately.
Overlap metrics, semantic similarity, retrieval quality, latency, and audit completeness are kept in separate views so answer similarity is not mistaken for deployment readiness.
| Model | ROUGE-L | BERT-F1 | Token F1 | Uni Prec | Bi Prec |
|---|
MSAS component families
Auditability is separated from answer similarity.
Clinical interpretation boundary
Overlap and semantic metrics are useful signals, but unsupported claims, unsafe omissions, missing provenance, and unresolved contradictions are tracked separately because answer similarity alone is not enough for deployment safety.
Reproducibility and release scope
The final build presents MedSwin as a reconstructable pipeline.
The release path covers data filters, teacher-label utilities, QLoRA scripts, merge specifications, retrieval policy packages, and evaluation harnesses.
Grounded clinical assistant
MedSwin returns evidence-linked answers rather than unsupported narratives.
The final system keeps patient context, guideline passages, sufficiency decisions, contradiction handling, and provenance together so clinicians can inspect how an answer was formed.
Return to surface