Advancing Detector Calibration and Event Reconstruction in Water Cherenkov Detectors through Differentiable Simulation

Next-generation neutrino experiments require unprecedented precision in event reconstruction and calibration, yet traditional decoupled analysis methods struggle to capture complex parameter correlations. We introduce a transformative approach using differentiable physics simulations that unifies calibration and reconstruction into a single optimization framework. By implementing modern automatic differentiation techniques through parallel computing frameworks, our method enables simultaneous optimization of coupled detector effects and physics parameters. This approach demonstrates enhanced reconstruction capabilities while maintaining full physics interpretability, showing promising results for next-generation experiments like Hyper-Kamiokande, where precise calibration is crucial for advancing our understanding of neutrino physics.