Speaker
Description
Femtoscopy provides a unique laboratory for studying hadron-hadron interactions via two-particle correlation functions (CFs), yet its predictive power is fundamentally limited by the unknown emission source $\mathcal{S}(r)$. We present a model-independent reconstruction of the proton-proton source by formulating the Koonin-Pratt equation as an inverse problem, stabilized through Tikhonov regularization. Applied to ALICE data in p--p collisions at $\sqrt{s}=7$~TeV and Pb--Pb collisions at $\sqrt{s_{\mathrm{NN}}}=5.02$~TeV across three centrality classes, our approach reveals a universal source-size dependence: compact sources exhibit pronounced non-Gaussian tails driven by the core-halo structure of resonance decays, whereas larger fireballs approach Gaussianity. The extracted sources are validated by their consistency with independent neural-network and transport-model calculations, and by their ability to correctly predict p--$\Lambda$ correlations without additional tuning. These results establish a data-driven, quantitative benchmark for source-shape uncertainties, laying the groundwork for extending the ``source-known'' paradigm to the extraction of novel hadron-hadron interactions.