Damavand: A Neural-Network Framework for Diffractive QCD Analyses

• Hadi Hashamipour (School of Particles and Accelerators, IPM)

In this talk, I will introduce **Damavand**, a neural-network-based QCD analysis framework developed for the determination of diffractive parton distribution functions (diffractive PDFs). Building upon the MontBlanc framework, Damavand combines perturbative QCD calculations at next-to-leading order (NLO) and next-to-next-to-leading order (NNLO) with DGLAP evolution, flexible neural-network parametrizations, and Monte Carlo replica methods for uncertainty estimation. As its first application, Damavand has been employed to extract the **MAP-diff-PDF1.0** distributions from HERA diffractive deep-inelastic scattering measurements within the fracture-function formalism. I will discuss the main features of the framework, its computational methodology, and selected phenomenological results, including comparisons with existing diffractive PDF determinations. Finally, I will briefly outline possible extensions and future