Lin Li

The paper introduces CanaryRAG, a novel dual-path runtime defense mechanism that detects RAG Knowledge Base Leakage attacks by embedding canary tokens into retrieved knowledge chunks.

The paper introduces FIRCE, a framework that enhances intrusion detection systems by combining conformal evaluation for uncertainty quantification and drift detection with an adaptive chunking mechanism to robustly handle concept drift in real-world network traffic.

The paper proposes FreeUp, a frequency-decoupled framework that improves encrypted network anomaly detection by separately modeling and fusing low- and high-frequency components of traffic data.

This paper analyzes the x402 agentic payment protocol, demonstrating through five concrete, practical attacks that it is vulnerable across multiple stages of its payment workflow.

This paper introduces a novel class of backdoor attacks that exploit the numerical side effects of LLM inference optimization, achieving high success rates while maintaining clean accuracy.

The paper proposes VDSB-GWSyn, a Diffusion Schrödinger Bridge framework, to synthesize controllable and anatomically feasible guidewire images on coronary angiography (CAG) scans, significantly improving the performance of guidewire endpoint localization for robot-assisted PCI.

The paper introduces Loong, a novel human-like agent that significantly improves long document translation by adaptively selecting and utilizing optimal historical context using a specialized memory module and reinforcement learning.

The paper introduces MechVQA, a comprehensive dataset and benchmark for mechanical drawing understanding, and proposes the MechVL model, which significantly improves Multimodal LLMs' performance on these specialized tasks.

AdaCodec introduces a predictive visual coding scheme for video MLLMs, significantly improving efficiency and performance by transmitting only inter-frame changes and full reference frames when necessary.

The paper introduces Moment-Video, a new benchmark that diagnoses the ability of video MLLMs to understand brief, critical visual events, revealing that current models struggle significantly with temporal fidelity.

The paper introduces Tree-like Self-Play (TSP), a novel framework that treats secure code generation as a fine-grained decision process, significantly improving LLM security by forcing the model to self-correct localized vulnerabilities.

The paper introduces and analyzes cross-session stored prompt injection, demonstrating that persistent system state transforms prompt injection from a temporary model-level threat into a long-lived, system-level vulnerability in agentic systems.

The paper introduces RedEdit, an agentic red-teaming framework that demonstrates that malicious images can be easily edited to bypass safety classifiers while retaining their harmful semantics.