Analog Front-End Flow Optimization
An AI-EDA framework for automated analog amplifier design and closed-loop optimization.
This project explores an AI-EDA framework for automated analog amplifier design, integrating LLMs and GNNs across specification understanding, behavior-level topology planning, transistor-level netlist generation, and closed-loop parameter optimization.
It proposes a metric-conditioned graph neural network for joint prediction of gain, GBW, phase margin, and power consumption, then uses the prediction results to guide structure and parameter optimization.