Qi Wang

The paper proposes a tool-mediated LLM architecture for autonomous cyber defense, formally proving its stability and demonstrating that it significantly reduces an attacker's expected payoff in real-world attack graph simulations.

PhishSigma++ is a novel entity-relation-based detector that improves malicious email detection by focusing on invariant functional relationships between typed entities, significantly outperforming text-centric models under adversarial manipulation.

The paper proposes Ellipsoid Control, a white-list defense mechanism that uses benign data geometry to constrain model updates, thereby enhancing jailbreak safety while preserving the utility of harmless inputs.

The paper proposes an unsupervised bi-level adversarial training framework to enhance LLM safety steering, achieving strong zero-shot defense against unseen and evolving jailbreak prompts.

The paper proposes formulating RAG design as an architecture search problem and introduces RAISE, a comprehensive framework and benchmark for systematically optimizing RAG hyperparameters.

The paper introduces Semantic Triplet Restoration (STR), a novel protocol that converts complex table structures into atomic semantic triplets, improving table question answering by providing explicit semantic context and reducing reliance on layout-dependent serializations.

The paper introduces VIABLE, the first benchmark for evaluating Vision-Language Models (VLMs) as judges for Visually Impaired Assistance (VIA), finding that current models are largely unreliable and proposing VIA-Judge-Agent to improve evaluation.

The paper introduces Group Prioritized Off-Policy Optimization (POPO), a novel framework that efficiently accelerates RL finetuning for LLM reasoning by leveraging effective off-policy training batches without requiring costly additional data rollouts.

The paper proposes a novel Generative Counterfactual Attention-guided Network (GCAN) that uses multimodal connectomes and brain atlas knowledge to provide explainable and highly accurate diagnosis of cognitive decline.

The paper proposes PaW, a co-training framework that uses standard RL rollouts to provide auxiliary world model supervision directly during policy training, significantly improving language agent performance.

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.

This survey provides a systematic framework and taxonomy for evidence tracing and execution provenance in LLM agents, addressing the difficulty of verifying and auditing complex agent behaviors.

The paper proposes OneReason, a framework that enhances the reasoning capability of generative recommendation models by focusing on improving item perception and structuring user behavior into coherent latent interests.