Hyeonjeong Ha

contrastive learning (1) robust optimization (1) adversarial robustness (1) image generation (1) curriculum learning (1) model robustness (1) self-supervised learning (1) neural architecture search (1) adversarial training (1) adversarial attack (1) adversarial perturbation (1) text-to-image model (1) multimodal large language model (1) retrieval-augmented generation (1) non-contrastive learning (1) zero-shot na (1) affordance composition (1) concept design (1) targeted adversarial attack (1) zero-shot proxy (1)