Tianwei Zhang

AudioTrust: Benchmarking The Multifaceted Trustworthiness of Audio Large Language Models

EchoBench: Benchmarking Sycophancy in Medical Large Vision-Language Models

Image Can Bring Your Memory Back: A Novel Multi-Modal Guided Attack against Image Generation Model Unlearning

Autoregressive-based Progressive Coding for Ultra-Low Bitrate Image Compression

SafeSearch: Automated Red-Teaming for the Safety of LLM-Based Search Agents

Work Zones challenge VLM Trajectory Planning: Toward Mitigation and Robust Autonomous Driving

Turning Bias into Bugs: Bandit-Guided Style Manipulation Attacks on LLM Judges

Picky LLMs and Unreliable RMs: An Empirical Study on Safety Alignment after Instruction Tuning

DSA: Efficient Inference For Video Generation Models via Distributed Sparse Attention

Warfare: Breaking the Watermark Protection of AI-Generated Content

FaceID-6M: A Large-Scale, Open-Source FaceID Customization Dataset

Turn That Frown Upside Down: FaceID Customization via Cross-Training Data

BitHydra: Towards Bit-flip Inference Cost Attack against Large Language Models

Mask Image Watermarking

Towards Resilient Safety-driven Unlearning for Diffusion Models against Downstream Fine-tuning

BSemiFL: Semi-supervised Federated Learning via a Bayesian Approach

Taught Well Learned Ill: Towards Distillation-conditional Backdoor Attack

Safe + Safe = Unsafe? Exploring How Safe Images Can Be Exploited to Jailbreak Large Vision-Language Models

Transstratal Adversarial Attack: Compromising Multi-Layered Defenses in Text-to-Image Models

TRUST-VLM: Thorough Red-Teaming for Uncovering Safety Threats in Vision-Language Models

TREANT: Red-teaming Text-to-Image Models with Tree-based Semantic Transformations

Impact-driven Context Filtering For Cross-file Code Completion

VideoShield: Regulating Diffusion-based Video Generation Models via Watermarking

An Engorgio Prompt Makes Large Language Model Babble on

REPOFILTER: Adaptive Retrieval Context Trimming for Repository-Level Code Completion

A Benchmark for Semantic Sensitive Information in LLMs Outputs

Warfare: Breaking the Watermark Protection of AI-Generated Content

Cowpox: Towards the Immunity of VLM-based Multi-Agent Systems

PRIME: Protect Your Videos From Malicious Editing

Towards False-claim-resistant Model Ownership Verification via Targeted Fingerprint

SuperMark: Robust and Training-free Image Watermarking via Diffusion-based Super-Resolution