Profile OmniFold Incorporates Nuisance Parameters For Unfolding
Researchers introduce Profile OmniFold (Dec 8, 2025), an extension of the OmniFold classifier-based EM unfolding algorithm that profiles nuisance parameters. The paper demonstrates the method on a Gaussian toy example and simulated CMS data from the Large Hadron Collider, showing how to propagate detector-model uncertainties into unfolded high-dimensional distributions. This enables more robust uncertainty estimates in simulation-based unfolding workflows.
Key Points
- 1Extends the OmniFold algorithm to profile nuisance parameters during simulation-based unfolding tasks
- 2Enables principled uncertainty propagation when the detector forward model is only approximately specified
- 3Allows practitioners to produce unfolded distributions with calibrated systematic uncertainties for LHC analyses
Scoring Rationale
Methodological advance enabling uncertainty-aware unfolding, but limited to simulation-based detector problems and pending wider validation.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
