This work presents a software-only Paton Assist overlay that applies a viability (admissibility) gate to reconstructed detector events by enforcing cross-layer consistency between external, intermediate, and internal measurement references. As detector segmentation and machine-learning–assisted reconstruction increase ambiguity and boundary effects, the overlay improves robustness, interpretability, and long-term stability of reconstruction pipelines without modifying detector hardware, sensing physics, or existing reconstruction algorithms.
Andrew John Paton (Fri,) studied this question.