Title: Leveraging Latent Symmetries for Target Localization, Identification, and Sensing

Abstract: This study demonstrates that scatterer arrays engineered with latent ("hidden") symmetries can function effectively as sensors. Utilizing the capacitance matrix as a standard model for three-dimensional hybridization, we investigate how the introduction of an "intruder" scatterer disrupts these latent symmetries. By evaluating the extent to which each symmetry is compromised, we are able to determine the intruder’s position and radius. While a dictionary-based method can accomplish this task, our results indicate that Bayesian inference and artificial neural networks (specifically multi-layer perceptrons) offer superior performance when measurement noise is present. To the best of our knowledge, this represents the first successful application of latent symmetries to sensing challenges. Furthermore, this work marks the inaugural observation of latent symmetries within a three-dimensional open system that defies approximation by sparse graphs.

Source: arXiv Generated at: 2026-06-02 00:00:00 UTC