Focuses on generative AI and robust learning: continual and safe adaptation of diffusion models; cross-modal alignment for vision–language; invariant/equivariant representations for multi-source dataset shift and group robustness; and scalable Bayesian uncertainty for deployment.