Amélie Chatelain presents a new method called NEWMA (No prior knowledge Exponentially Weighted Moving Average) to detect conformational changes in molecular dynamics simulations. NEWMA uses optical or CPU-based random features to analyze trajectories without prior knowledge. It was shown to outperform existing methods like diffusion maps in detecting changes in SARS-CoV-2 simulations, with optical random features providing faster computation. Future work includes using reinforcement learning to guide molecular dynamics simulations.