PIDT: Physics-Informed Digital Twin for Optical Fiber Parameter Estimation

This work addresses parameter estimation in optical fibers, which is incremental as it builds on existing neural operator methods.

The authors tackled the problem of optical fiber parameter estimation by proposing a physics-informed digital twin (PIDT) that combines a parameterized split-step method with a physics-informed loss, resulting in improved accuracy and convergence speed with lower complexity compared to previous neural operators.