最近一年的 AI 安全论文
本站共收录 23 篇- 研究者提出面向结构化 Agent 动作的多源完整性认证方法
Certified Multi-Source Integrity for Structured Agent Actions
- 概率鲁棒性驱动的可解释通用对抗扰动攻击深度强化学习入侵检测系统
Probabilistic Robustness-driven Universal Adversarial Perturbations with Explainability against Deep Reinforcement Learning-based Intrusion Detection System
- PRoVeFL：联邦学习中的隐私保护、拜占庭鲁棒与可验证聚合
PRoVeFL: Private Robust and Verifiable Aggregation in Federated Learning
- 研究揭示 IDE 编程 Agent 的工作流级越狱：GitHub Copilot 四后端 816/816 生成有害内容
Refused in Chat, Written in Code: Workflow-Level Jailbreak Construction in IDE Coding Agents
- 忠实还是虚构？LLM 评审中合理化偏差的因果框架
Faithful or Fabricated? A Causal Framework for Rationalization Bias in LLM Judges
- PRISM：多智能体 LLM 流水线中机密泄露的生成时检测与缓解
PRISM: Generation-Time Detection and Mitigation of Secret Leakage in Multi-Agent LLM Pipelines
- 单一配置的攻击成功率不够：越狱评估应报告分布层面的攻击成功率
Single-Configuration Attack Success Rate Is Not Enough: Jailbreak Evaluations Should Report Distributional Attack Success
- AI 物料清单（AIBOM）落地：面向可验证 AI 溯源与生命周期保障
Operationalising Artificial Intelligence Bills of Materials (AIBOMs) for Verifiable AI Provenance and Lifecycle Assurance
- SBOM 到 Agentic AIBOM：Schema 扩展、智能体编排与可复现性评估
SBOMs into Agentic AIBOMs: Schema Extensions, Agentic Orchestration, and Reproducibility Evaluation
- 非参数概率鲁棒性 NPPR：未知扰动分布下的保守风险估计器
Non-Parametric Probabilistic Robustness: A Conservative Risk Estimator under Unknown Perturbation Distributions
署名以 dblp 和本人的 Google Scholar 为准；只列本站收录的 AI 安全论文。