Haoran Zheng

retrieval-augmented generation (2) contrastive learning (1) semi-supervised learning (1) few-shot learning (1) optimal transport (1) knowledge distillation (1) multimodal learning (1) vision-language model (1) positional encoding (1) graph edit distance (1) membership inference attack (1) text clustering (1) safety awareness (1) maximum common subgraph (1) adversarial poisoning (1) graph neural network (1) graph similarity (1) harmful output (1) unsafe content (1) privacy attack (1)