Ke Li

The paper introduces SADBench, a systematic benchmark designed to evaluate both the effectiveness of steganographic attacks injecting harmful content and the robustness of steganalysis defenses against these threats.

The paper systematically measures the risk of current image-to-3D models generating harmful geometries, finding that these models are effective at reconstruction and existing safeguards are insufficient.

This paper measures the prevalence of recurring vulnerability patterns (variants) across multiple AI infrastructure repositories and proposes INFRASCOPE, a framework to automatically detect these variants.

The paper introduces Causal Editing (CODE), a new paradigm that improves knowledge updates in LLMs by grounding fact injection in causal narratives, drastically reducing self-refutation rates.

The paper introduces a new dataset (SHDF) and a framework (T-AVFD) to robustly detect audio-visual deepfakes, specifically addressing the challenge posed by singing vocalizations.

The paper proposes a dual-interventional framework to characterize how linguistic structures and contextual cues influence LLMs' spatial reasoning for navigation, finding that topological information is crucial, while semantic details can be unreliable.

The eMoT framework enhances multi-step reasoning in LLMs by treating reasoning as an evolving memory, stabilizing performance through symbolic computation and structured refinement.