Huili Wang

Privacy (2) Differential Privacy (1)

membership inference attack (3) retrieval-augmented generation (2) privacy attack (1) model robustness (1) knowledge distillation (1) multimodal learning (1) gradient aggregation (1) adversarial attack (1) black-box attack (1) poisoning attack (1) mixture of expert (1) vision-language model (1) multimodal large language model (1) temporal dependency (1) privacy protection (1) model update (1) gradient perturbation (1) safety awareness (1) ai-generated text (1) adversarial poisoning (1)