arXiv subjects

Wentai Deng

Publications and source records attributed to Wentai Deng.

3 recordsLinked to original sources

arXiv2025

Minimizing False Positives in Static Bug Detection via LLM-Enhanced Path Feasibility Analysis

Static bug analyzers play a crucial role in ensuring software quality. However, existing analyzers for bug detection in large codebases often suffer from high false positive rates. This is primarily due to the limited capabilities of analyzers in path feasibility validation with multiple conditional branches and complex data dependencies. While current LLM-based approaches attempt to address this issue, their effectiveness remains limited due to insufficient constraint cascade analysis and scalability challenges in large projects. To address this challenge, we propose an iterative path feasibility analysis framework LLM4PFA. By leveraging LLM agent based targeted constraint reasoning, and key context-aware analysis driven by agent planning, LLM4PFA effectively enhances complex inter-procedural path feasibility analysis for minimizing false positives in static bug detection. Evaluation results show that LLM4PFA precisely filters out 72% to 96% false positives reported during static bug detection, significantly outperforming all the baselines by 41.1% - 105.7% improvements; meanwhile LLM4PFA only misses 3 real bugs of 45 true positives.

cs.SE↗

arXiv2024

Vul-RAG: Enhancing LLM-based Vulnerability Detection via Knowledge-level RAG

Although LLMs have shown promising potential in vulnerability detection, this study reveals their limitations in distinguishing between vulnerable and similar-but-benign patched code (only 0.06 - 0.14 accuracy). It shows that LLMs struggle to capture the root causes of vulnerabilities during vulnerability detection. To address this challenge, we propose enhancing LLMs with multi-dimensional vulnerability knowledge distilled from historical vulnerabilities and fixes. We design a novel knowledge-level Retrieval-Augmented Generation framework Vul-RAG, which improves LLMs with an accuracy increase of 16% - 24% in identifying vulnerable and patched code. Additionally, vulnerability knowledge generated by Vul-RAG can further (1) serve as high-quality explanations to improve manual detection accuracy (from 60% to 77%), and (2) detect 10 previously-unknown bugs in the recent Linux kernel release with 6 assigned CVEs.

cs.SE↗

arXiv2023

Leveraging static quantum many-body scars into period-doubled responses

We propose a scheme that generates period-doubled responses via periodically driving certain Hamiltonians hosting quantum many-body scars, akin to recent experimental observations in driven Rydberg atom arrays. Our construction takes advantage of an su(2) spectrum generating algebra associated with the static quantum-scarred Hamiltonian, which enacts a $π$-rotation within the scar subspace after one period of time evolution with appropriately chosen driving parameters. This yields period-doubled (subharmonic) responses in local observables for any choice of initial state residing in the scar subspace. The quasienergy spectrum features atypical $π$-paired eigenstates embedded in an otherwise fully thermal spectrum.The protocol requires neither a large driving frequency nor a large driving amplitude, and is thus distinct from the prethermalization physics in previous investigations of the driven PXP model. We demonstrate our scheme using several spin-1/2 and spin-1 quantum scarred models possessing an exact su(2) spectrum generating algebra, as well as a symmetry-deformed PXP model, where the su(2) algebra is only approximate. Our results extend the class of models hosting quantum many-body scars that can be leveraged to yield time-crystalline behaviors under periodic driving.

cond-mat.stat-mech↗