Circuit Design through Large Language Models and Reinforcement Learning

Irene Terpstra · September 20, 2023

MEng

I recently graduated with my Master’s from MIT. I worked with the MIT IBM Watson AI lab and MIT Energy-Efficient Circuits Lab. My project focused on using Large Language Models (LLMs) to improve the design process of Analog Integrated Circuits. In my thesis, I propose a method that uses the power and knowledge of LLMs to guide the training of a reinforcement learning algorithm.

Get the PDF here.