Driving Epidemic Models with AI Agents: the Epydemix Agent Framework
Epydemix Agent Framework is an additive layer enabling LLM agents to configure, validate, execute, and inspect stochastic epidemic models without custom code.
Why this signal matters
The Epydemix Agent Framework provides an additive layer over Epydemix, an open-source Python library for stochastic compartmental epidemic modeling. Designed to improve the reliability of LLM agents interacting with scientific software, the framework equips agents with four core capabilities: model and parameter discovery, preventive validation of declarative scenario specifications, execution using tested library code, and result inspectability. Agents can manage the full lifecycle—from natural language scenario input to quantitative outputs, plots, and interpretations—without generating custom code. Each workflow step uses input files and outputs separate, auditable bundles to ensure reproducibility. In an evaluation across 50 sessions and five modeling tasks comparing the framework to direct Python interface usage, the framework lowered interaction turns, output tokens, and cost on most tasks, with occasional resource trade-offs when prioritizing per-point reproducibility.
Actionable summary
Epydemix Agent Framework augments the Epydemix Python library with model/parameter discovery, declarative scenario validation, audited execution bundles, and result inspectability, reducing agent turns and token overhead.
- Agent usefulness
- 80/100
- Confidence
- 90%
- Canonical data
- JSON + Markdown
What builders should check
- Review paper arXiv:2609.28692v1 for architectural patterns on declarative scenario validation and auditable agent execution bundles in scientific modeling.
- Evaluate Epydemix Agent Framework when building agent interfaces for stochastic compartmental epidemic simulations.