My current research focuses on the safety and security of language-model agents. Earlier work spans generative AI systems and intelligent healthcare.
arXiv preprint arXiv:2605.29224
AgentREVEAL diagnoses retrieval-induced safety degradation, while HarmURLBench evaluates relevance-triggered failure modes across 1,405 URLs and 320 harmful behaviours.
IEEE Transactions on Machine Learning in Communications and Networking
An IoMT-based incremental-learning framework for smart healthcare diagnosis, improving anomaly detection and classification while reducing runtime versus batch-learning baselines.
ACM Transactions on Multimedia Computing, Communications and Applications
A survey of modern generative image-generation systems across quality, diversity, interpretability, computational efficiency, and practical deployment.