Hybrid physics-informed data-driven models outperform both pure physics and pure machine-learning models in generalizability and data efficiency across three photonic system benchmarks.
Inverse design of discrete Raman amplifiers using an invertible neural network for ultra-wideband optical transmis- sion based on hollow core fibers,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.optics 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Beyond white- and black-box modeling tools in optical communications and optical computing: physics-informed data-driven modeling
Hybrid physics-informed data-driven models outperform both pure physics and pure machine-learning models in generalizability and data efficiency across three photonic system benchmarks.