LLM orchestrates design of silicon-on-insulator directional couplers

Researchers have developed a Large Language Model (LLM) agent capable of orchestrating the design of silicon-on-insulator (SOI) directional couplers. This agent proposes geometrical dimensions and judges convergence, while relying on deterministic solvers for the actual physics calculations. A frequency-domain eigenmode solver determines the coupling coefficient, and a Finite-Difference Time-Domain (FDTD) stage validates the design, ensuring consistency within a 2D effective-index model. The LLM successfully delivered a 50/50 splitter with a minimal residual error. AI

IMPACT Demonstrates LLMs' potential to automate complex scientific design tasks, potentially accelerating innovation in optics and photonics.

RANK_REASON The cluster contains an academic paper detailing a novel application of LLMs in a scientific design process. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• 2D computer graphics
• eigenmode solver
• finite-difference time-domain method
• large language model
• power dividers and directional couplers
• silicon on insulator

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →