CMA-Unfold A Covariance Matrix Adaptation unfolding algorithm for stacked calorimeter detectors

TL;DR

This paper introduces CMA-Unfold, an open-source algorithm based on CMA-ES, for reconstructing photon energy spectra from stacking calorimeter data, effectively handling noise and uncertainties in high-energy physics diagnostics.

Contribution

The paper presents a novel, robust unfolding algorithm that directly reconstructs spectra from depth-dose profiles without restrictive assumptions, improving analysis of calorimeter data.

Findings

Accurately recovers complex spectral shapes.

Tolerates percent-level deviations in detector layers.

Provides a flexible, noise-resilient analysis tool.

Abstract

Stacking calorimeters also refered as bremsstrahlung cannons widely used in inertial confinement fusion and ultra-intense laser plasma experiments have become essential diagnostics for characterizing short bursts of high-energy photons and charged particles. Extracting the underlying energy spectrum from these detectors requires solving an ill-posed inverse problem, often complicated by noise, secondary particle contamination, and uncertainties in the detector response. In this work, we introduce an open-source unfolding framework available under the username ggfauvel under CMA-unfold based on the Covariance Matrix Adaptation Evolution Strategy (CMA-ES), designed to reconstruct photon spectra directly from depth-dose profiles without imposing restrictive parametric assumptions. The algorithm demonstrates high robustness, accurately recovering complex spectral shapes while tolerating percent-level deviations in individual detector layers. This approach provides a flexible and noise-resilient tool for the analysis of stacking calorimeter data, with particular relevance for bremsstrahlung diagnostics in inertial confinement fusion and high-intensity laser applications.