CleanCLIP: Mitigating Data Poisoning Attacks in Multimodal Contrastive Learning

Open in new window