New Paradigm for AI: Reinforcement learning for EDA optimization, anomaly detection, generative AI

Unlike the development of models, we think that a new paradigm for AI is the complex application for open-ended problem.  In agreement with the new paradigm for AI, reinforcement learning for EDA optimization, anomaly detection, and generative AI are our research topics at time time. 

Firstly, in the reinforcement learning for EDA optimization, the parameter optimization for complex circuit is considered. To overcome handcraft simulation for circuit optimization, the trained policy networks and value networks under a reinforcement framework can provide automatic solution for circuit optimization, which can automate the circuit design process. 

Secondly, in the anomaly detection, the anomaly parts can be detected with high accuracy, which can be attractive in commercial fields.

Finally, the changeable and unchangeable parts can be mixed in the output of generative AI models; there is no explicit solution until now.