AI-Based Algorithms in Analog Circuit Topology Generation and Beyond

The automation of analog circuit design is rapidly evolving with recent advancement of AI/ML, especially large language models (LLMs). The demand for new algorithms in analog circuit design is driven by the need to handle increasing complexity and stringent design constraints in advanced technology nodes. Traditional design methods, which rely heavily on manual iterations, are often time-consuming and limited in their ability to explore the full design space. This talk will explore how recent advances in AI/ML are transforming the analog EDA landscape, enabling more efficient and innovative design processes. We will discuss recent breakthroughs in ML-driven approaches for analog circuit design, including topology generation, performance optimization, and design space exploration. Examples will be provided from recent developments in general analog circuit design, with extensions into RF circuits and high-frequency applications.