Bo Wang

The paper introduces FEDB, a novel confidential database design that eliminates cryptographic operations from the critical query path, significantly reducing performance overhead for secure querying over sensitive data.

MemHint is a neuro-symbolic static analysis pipeline that significantly improves memory leak detection in C/C++ by combining LLM semantic understanding with Z3 symbolic reasoning, detecting more leaks than existing tools.

The paper introduces MyPhoneBench, a new framework that demonstrates that current phone-use agents often fail to respect user privacy, even when successfully completing simple tasks, primarily due to unnecessary data disclosure.

The paper introduces the Black-Hole Attack, a poisoning vulnerability that exploits geometric defects in high-dimensional embedding spaces to force malicious vectors into the top-k results of vector database queries.

This paper systematically investigates unlearnable examples (UEs) across diverse training paradigms, finding that existing UEs fail under pretraining-finetuning (PF) settings, and proposes Shallow Semantic Camouflage (SSC) to maintain unlearnability.

This paper introduces a novel denial-of-service attack targeting multi-round transaction simulation by exploiting inter-transaction dependencies within smart-contract state.

This paper reinterprets catastrophic overfitting (CO) in Fast Adversarial Training (FAT) as a weak backdoor mechanism, proposing backdoor-inspired strategies to mitigate this generalization failure.

The paper proposes a Distribution-aware Dynamic Guidance (DDG) strategy to mitigate catastrophic overfitting and the robustness-accuracy trade-off inherent in Fast Adversarial Training (FAT) by dynamically adjusting perturbation budgets and supervision signals based on sample confidence.

The paper introduces LoopTrap, an automated red-teaming framework that demonstrates how malicious prompts can poison the termination judgment of LLM agents, causing unbounded computation.

RouteScan introduces a non-intrusive framework that audits the safety of Mixture-of-Experts (MoE) LLMs by analyzing low-level GPU expert routing telemetry, achieving high accuracy even on unseen harmful prompts.

The paper introduces RA-MoE, a novel fine-tuning framework that leverages the internal routing structure of Mixture-of-Experts (MoE) models to improve performance on multilingual downstream tasks by aligning target-language routing patterns with English task-expert activations.

LoRA-Key introduces a user-centric watermarking framework that attaches a recoverable ownership key to LoRA modules via a standalone Watermark LoRA, providing lightweight, plug-and-play copyright protection without requiring per-LoRA retraining.

The paper introduces SAVE, a framework that uses on-policy feedback and the value function to self-supervise and improve reward models, significantly enhancing RLHF performance across multiple benchmarks.

The paper proposes MERIT, a dual-level, multi-horizon memory retrieval framework that significantly improves the performance of interactive text-to-SQL agents by providing both global and local memory guidance.