Low-temperature plasma adaptation in the course of machine learning controls of plasma medicine

This study unveils the optimization of chemical dosing in cold atmospheric plasma (CAP) treatments for biomedical applications under machine learning control. It is demonstrated that machine learning can guide CAP treatments to achieve precise cell viability outcomes even without using any explicit information about the plasma state, but implicitly, the chemical composition/dose of the plasma is optimized by the control. This work thus opens the “black box” of AI-controlled plasma treatment by revealing how treatment efficacy is linked to optimized chemical delivery. The proposed approach offers a framework for enhancing the precision, transparency, and adaptability of plasma-based therapies through intelligent control.