Back to Machine Learning for Science (ML4SCI)
GSoC 2026

Agentic Lagrangian Extraction from the Literature ML4SCI – HEPSIM5

Hundreds of BSM Lagrangians have been proposed in the literature, but translating them into validated FeynRules .fr model files is still a manual, time-consuming process. This project builds an agentic system inside the HEPTAPOD framework that automates this workflow. Given a BSM scenario description, the agent searches INSPIRE and arXiv, extracts Lagrangian terms into a typed schema, generates a syntactically correct .fr file using deterministic Jinja2 templates, and validates the output through HEPTAPOD’s FeynRulesToUFOTool and a downstream MadGraph smoke test. Every literature source, extraction decision, and validation outcome is recorded in an audit trail.

Project details

Contributor

kenwu

Mentors

Not available

Technologies

Not listed in the archive