Xinrong Hu

representation learning (1) adversarial robustness (1) speech synthesis (1) dialogue generation (1) narrative generation (1) deep learning (1) adversarial attack (1) recommendation system (1) computed tomography (1) masked language modeling (1) graph contrastive learning (1) low-resource setting (1) knowledge base question answering (1) contrastive regularization (1) input perturbation (1) subspace alignment (1) small object detection (1) voice cloning (1) text to speech (1) multi-view alignment (1)