The recent advancements in the research area of circuit design and optimization have seen a significant shift towards leveraging artificial intelligence (AI) and large language models (LLMs) to enhance efficiency and accuracy. Innovations in approximate circuit error computation using SAT and message-passing algorithms have enabled exact metrics computation, which was previously unattainable. Timing-driven synthesis methods have been improved by integrating metaheuristic optimization techniques, significantly reducing critical path delays. Data-driven approaches in electrical machine design have revolutionized preliminary design processes by incorporating AI-guided databases, leading to faster and more efficient designs. LLM-enhanced design space reduction and optimization for analog circuits have shown superior performance and generalizability, outperforming traditional methods. Additionally, the introduction of knowledge graph-based datasets for AMS circuit auto-design has addressed the issue of model hallucination, ensuring robust design generation. Finally, physics-informed LLM-agents for automated modulation design in power electronics systems have demonstrated remarkable improvements in design efficiency and error reduction, making them highly competitive in the field.

Noteworthy papers include one that proposes a framework for exact computation of error metrics in approximate circuits using SAT and message-passing algorithms, and another that introduces an LLM-based, physics-informed autonomous agent for automated modulation design, significantly reducing design time and error.