rayid-mojumder/Physics-inspired-ML-dataset-for-Photonic-Waveguide-Design: v1.0.0

🚀 Release v1.0.0 Preview We're excited to unveil v1.0.0 of Physics-inspired-ML-dataset-for-Photonic-Waveguide-Design. This release offers two complementary synthetic datasets tailored for cylindrical waveguides and on‐chip photonic interconnects in glass substrates—perfect for integrated photonics, HBM interposers, AI/GPU accelerators, and beyond. 📈 Available Datasets Physics-Only Dataset 50 000 samples generated directly from first‐principles waveguide equations Computes normalized frequency (V), mode‐field diameter (MFD), confinement (Γ), attenuation (α), effective index (nₑff), polarization ratios, cross‐coupling, and output power Ideal for pure physics‐inspired ML benchmarks in photonic device design Realistic Dataset (Noise + Experimental Infusion) Builds on the physics core, then injects 5 % Gaussian noise to emulate fabrication/measurement variability Calibrated against literature measurements (propagation loss, MFD, etc.) across fused silica, phosphate, chalcogenide, and doped‐glass waveguides Marries analytical rigor with real‐world data for robust, generalizable model training 🔧 Highlights Physics-Inspired FoundationsMarcuse's empirical MFD formula, Rayleigh scattering theory, Gaussian overlap coupling, exponential loss decay Noise InjectionControlled 5 % Gaussian perturbations to simulate device and measurement uncertainties Experimental Data InfusionTuned to match published propagation‐loss and mode‐field statistics for glass and silicon-based waveguides Target Applications Integrated Photonics High-Bandwidth Memory (HBM) Interposers AI/GPU Photonic Accelerators Optical Neural Networks & On-Chip Optical Links 📂 Quick Start git clone https://github.com/rayid-mojumder/Physics-inspired-ML-dataset-for-Photonic-Waveguide-Design.git cd Physics-inspired-ML-dataset-for-Photonic-Waveguide-Design