Yanting Wang

adversarial attack (2) adversarial defense (2) backdoor attack (1) agent security (1) security evaluation (1) prompt injection (1) document-level relation extraction (1) llm agent (1) defense mechanism (1) data-free detection (1) security benchmark (1) logical rule (1) multi-modal model (1) trojan detection (1) large language model (1) model context protocol (1) neural-logic framework (1) multi-modal input (1) certified robustness (1) tool poisoning (1)