Yongyuan Liang

robot manipulation (2) adversarial robustness (1) few-shot learning (1) model security (1) data poisoning (1) policy optimization (1) direct preference optimization (1) multimodal learning (1) visual grounding (1) robust reinforcement learning (1) reinforcement learning from human feedback (1) backdoor attack (1) adversarial attack (1) diffusion model (1) adversarial perturbation (1) poisoning attack (1) foundation model (1) policy network (1) behavior embedding (1) trajectory information (1)