Research map / Large language models / Agents, code and math
Automatic Construction of Clinical Scoring Systems with LLM Agents
Silas Ruhrberg Estévez, Christopher Chiu, Mihaela van der Schaar
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add constrained proposal-validation loop with tools instead of unconstrained LLM generation (ablation) | AgentScore · Mean AUROC over 8 clinical tasks | AUROC: 0.59 -> 0.71 | Improved |
| add iterative refinement (ablation) | AgentScore · Mean AUROC over 8 clinical tasks | AUROC: 0.69 -> 0.71 | Improved |
| add Jaccard redundancy filtering (ablation) | AgentScore · Mean AUROC over 8 clinical tasks | AUROC: 0.69 -> 0.71 | Improved |
| add rule-type diversity enforcement (ablation) | AgentScore · Mean AUROC over 8 clinical tasks | AUROC: 0.70 -> 0.71 | Improved |
| compare AgentScore against RiskSLIM and FasterRisk integer scores | AgentScore · 8 MIMIC-IV/eICU tasks | AUROC: +0.05 versus RiskSLIM and +0.03 versus FasterRisk | Improved |
| compare AgentScore unit-weighted checklists against unconstrained logistic regression, decision tree and AutoScore | AgentScore · Mean AUROC over 8 tasks | AUROC: 0.71 vs 0.73, 0.72, 0.72 | Hurt |
| compare AgentScore with Optimal Checklists on fixed features | AgentScore · Mean AUROC over 8 tasks | AUROC: 0.62 -> 0.71 | Improved |
Closest papers by meaning
- MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making (NeurIPS 2024 · similarity 0.87)
- GEMA-Score: Granular Explainable Multi-Agent Scoring Framework for Radiology Report Evaluation (AAAI 2026 · similarity 0.86)
- MedChain: Bridging the Gap Between LLM Agents and Clinical Practice with Interactive Sequence (NeurIPS 2025 · similarity 0.86)
- LLMs can construct powerful representations and streamline sample-efficient supervised learning (NeurIPS 2026 · similarity 0.85)
- ClinicalLab: Aligning Agents for Multi-Departmental Clinical Diagnostics in the Real World (NeurIPS 2025 · similarity 0.85)
- PhysicianBench: Evaluating LLM Agents in Real-World EHR Environments (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
