Hui Xue

Vulcan: Crafting Compact Class-Specific Vision Transformers For Edge Intelligence

PlugGuard: A Streaming Safeguard for Large Models via Latent Dynamics-Guided Risk Detection

Curvature-Aware Residual Prediction for Stable and Faithful Diffusion Transformer Acceleration Under Large Sampling Intervals

Thought Purity: A Defense Framework For Chain-of-Thought Attack

Strata-Sword: A Hierarchical Safety Evaluation towards LLMs based on Reasoning Complexity of Jailbreak Instructions

Dynamic Mixture of Curriculum LoRA Experts for Continual Multimodal Instruction Tuning

Towards the Resistance of Neural Network Fingerprinting to Fine-tuning

AIR: A Systematic Analysis of Annotations, Instructions, and Response Pairs in Preference Dataset

Modularized Self-Reflected Video Reasoner for Multimodal LLM with Application to Video Question Answering