Wei Yang Bryan Lim

AdaCache: Adaptive Caching and Context Augmentation for Efficient LLM Serving

Disrupting Hierarchical Reasoning: Adversarial Protection for Geographic Privacy in Multimodal Reasoning Models

FedSAGD: federated learning with stable and accelerated client gradient descent

AutoHete: An Automatic and Efficient Heterogeneous Training System for LLMs

Severing the Link: A Unified Adversarial Attack on Image and Video MLLMs via Generative Disruption

Conformal Regression under Distribution Shift: A Reinforcement Learning Method for Adaptive Uncertainty Quantification

Heterogeneous Federated PEFTs via Sparse Mixture-of-Experts

Q-CLIP: Unleashing the Power of Vision-Language Models for Video Quality Assessment through Unified Cross-Modal Adaptation

TrojanTools: Adaptive Indirect Prompt Injection on LLM Agents via Malicious Tool-Calling

ModelPirate: Security Analysis of Partial Merging Against Model Stealing Attacks

Verifiably Forgotten? Gradient Differences Still Enable Data Reconstruction in Federated Unlearning

FedRAM: Federated Reweighting and Aggregation for Multi-Task Learning

AegisGuard: RL-Guided Adapter Tuning for TEE-Based Efficient & Secure On-Device Inference

Enhancing Federated Domain Adaptation with Multi-Domain Prototype-Based Federated Fine-Tuning

HoLoRA: Combining Orthogonal Fine-Tuning and LoRA with Householder Reflectors