Jiacheng Liang

safety alignment (2) contrastive learning (1) adversarial robustness (1) self-supervised learning (1) efficient computing (1) harmful content (1) reinforcement learning from human feedback (1) model alignment (1) convolutional neural network (1) adversarial defense (1) text watermark (1) reward model (1) edge computing (1) watermark detection (1) jailbreak attack (1) adversarial prompt (1) red teaming (1) video processing (1) model acceleration (1) data curation (1)