18 papers
Preserving Provenance in Shared KV Caches for LLM Serving
Wei Song, Yuxin Cao, Xi Zheng +2
Production LLM serving stacks combine an inference engine's local prefix cache with a shared KV-cache tier for fleet-wide reuse. The local cache distinguishes requests by adapter,…
Rule-Constrained Assignment for Cue-Ball Identification in Broadcast Snooker
Yuxin Cao, Wei Song, Yuezhong Wu +1
Accurate cue-ball identification is essential for metric analysis of broadcast snooker. Existing systems evaluate each candidate independently against a fixed white prototype and r…
COVER: Codec-Robust Video Watermarking with Generative Video Priors
Yuxin Cao, Hao Yang, Ziqi Ding +2
Video watermarking underpins copyright protection and provenance for generated media, yet almost every video is compressed by a codec before it is stored or shared. A codec discard…
Beyond Patch Removal: Persistent Adversarial Effects in Vision-Language-Action Policies
Enhao Wu, Fusen Guo, Yuxin Cao +3
Adversarial patches to Vision-Language-Action (VLA) policies can cause both immediate action corruption and persistent state effects that remain after the patch is removed. Existin…
Runtime Safety Filtering for Two-Terminal Hazards in Robotic Battery Recycling
Yuxin Cao, Wei Song, Xianglin Yang +4
Runtime safety filters for learned manipulation policies typically define unsafe states as unions of object-wise keep-out regions. This representation can be unnecessarily restrict…
Do Input-Level Defenses Transfer to Observation-Level Attacks on VideoLLMs?
Bangshuo Zhu, Wei Song, Yuxin Cao +4
Video Large Language Models (VideoLLMs) are increasingly deployed in safety-critical applications such as content moderation and video analytics. To process long videos efficiently…
KeyBound: Keyed and Host-Bound Learned Audio Watermarking for Speech Provenance
Bangshuo Zhu, Yuxin Cao, Weifei Jin +4
Audio watermarking is a proactive route to attributing synthetic speech to its source. Learned audio watermarks are typically judged by payload recovery after a fixed catalog of si…
The Temporal Moderation Gap: Text-to-Video Safety Filters Are Blind to Harm in Motion
Yuxin Cao, Fusen Guo, Yuezhong Wu +2
Text-to-video (T2V) services inherit their safety stack from image generation, pairing a keyword prompt filter with a per-frame checker that blocks a clip whenever one sampled fram…
SpreadMark: Robust Image Watermarking via Spread-Spectrum Embedding
Wei Song, Yuxin Cao, Zhenchang Xing +4
Invisible image watermarks are increasingly used for deepfake detection and provenance tracking, where they must survive not only incidental distortions but also deliberate removal…
Caved or Convinced: Temporal Sampling Gates Claim Deference in Video Large Language Models
Yuxin Cao, Wei Song, Jingling Xue +1
When asked which of two events came first, video large language models can fail in two opposite ways: cave to a false claim, or reject a true one. Prior video sycophancy work measu…
Membership Inference Attacks Against Video Large Language Models
Wei Song, Yuxin Cao, Ziqi Ding +3
Video large language models (VideoLLMs) are increasingly trained or instruction-tuned on large-scale video--text corpora collected from heterogeneous sources, raising an immediate…
VideoSTF: Stress-Testing Output Repetition in Video Large Language Models
Yuxin Cao, Wei Song, Shangzhi Xu +2
Video Large Language Models (VideoLLMs) have achieved strong performance on video understanding tasks, yet existing benchmarks evaluate only what models predict, leaving the stabil…
FAIRT2V: Training-Free Debiasing for Text-to-Video Diffusion Models
Haonan Zhong, Wei Song, Tingxu Han +3
Text-to-video (T2V) diffusion models have achieved rapid progress, yet their demographic biases, particularly gender bias, remain largely unexplored. We present FairT2V, a training…
DeMark: A Query-Free Black-Box Attack on Deepfake Watermarking Defenses
Wei Song, Zhenchang Xing, Liming Zhu +2
The rapid proliferation of realistic deepfakes has raised urgent concerns over their misuse, motivating the use of defensive watermarks in synthetic images for reliable detection a…
DUAP: Dual-task Universal Adversarial Perturbations Against Voice Control Systems
Suyang Sun, Weifei Jin, Yuxin Cao +2
Modern Voice Control Systems (VCS) rely on the collaboration of Automatic Speech Recognition (ASR) and Speaker Recognition (SR) for secure interaction. However, prior adversarial a…
Poisoning Prompt-Guided Sampling in Video Large Language Models
Yuxin Cao, Wei Song, Jingling Xue +1
Video Large Language Models (VideoLLMs) are increasingly deployed as automated moderators on user-generated video platforms, where a few unwatched seconds of harmful footage are en…
Help or Hurdle? Rethinking Model Context Protocol-Augmented Large Language Models
Wei Song, Haonan Zhong, Ziqi Ding +2
The Model Context Protocol (MCP) enables large language models (LLMs) to access external resources on demand. While commonly assumed to enhance performance, how LLMs actually lever…
Failures to Surface Harmful Contents in Video Large Language Models
Yuxin Cao, Wei Song, Derui Wang +2
Video Large Language Models (VideoLLMs) are increasingly deployed on numerous critical applications, where users rely on auto-generated summaries while casually skimming the video…