Security of World-Model-Based Embodied AI: A Lifecycle of Threats, Defenses, and Evaluation
该论文是一篇关于基于世界模型的具身人工智能（Embodied AI）安全性的综述。世界模型为具身智能体提供预测核心：将观测压缩为状态，模拟基于行动的未来，并支持超越反应式控制的规划。然而，这一预测层引入了新的安全边界——攻击可以从数据、传感
2026-08-03 归档
今日共收录 797 条安全情报，包含 119 个 CVE，373 篇安全通告，以及 59 篇研究论文。
← 返回归档列表Secure Aggregation for Privacy-Preserving Federated Learning on Clinical EEG Data
本文提出一种针对临床脑电图（EEG）数据的隐私保护联邦学习框架，核心机制是基于掩码的安全聚合（secure aggregation）。联邦学习虽然允许多个机构在不交换原始临床EEG数据的前提下共同训练模型，但个体模型更新仍可能泄露患者隐私。
Technology-Enhanced Tabletop Exercises for Cybersecurity Education: Lessons Learned
该论文（类别 cs.CY）探讨将技术增强的桌面演练（TTX）引入网络安全课程教学，以弥补高校在团队协作、事件响应和 IT 治理等复杂场景训练上的不足。TTX 在专业领域已很成熟，但在大学中仍较少使用。作者基于 INJECT Exercise
Bridging the Gap Between PHE and FHE: A Performance and Trade-off Analysis of The Somewhat Homomorphic BGN Cryptosystem
本文针对同态加密（HE）在实际应用中面临的权衡问题——轻量级部分同态加密（PHE）计算高效但功能有限，而全同态加密（FHE）功能强大但计算开销巨大——提出并分析了介于两者之间的 Boneh-Goh-Nissim（BGN）密码系统。BGN 属
Checking Information Flow in Cloud-based IoT Access Control Policies (Extended Version)
本文研究云环境下物联网（IoT）访问控制策略中的信息流安全问题。许多云服务商（如AWS IoT Core）提供访问控制机制，但其配置不当可能导致严重安全漏洞。作者指出，在设备具有不同信任级别或隶属于不同子系统的场景中，仅独立验证单个权限是否
Benign on Label, Malicious by Design: Clean-Label Dormant-to-Activated Backdoor via Machine Unlearning with Removable Camouflage
本文提出一种新型的干净标签休眠后门攻击框架，其核心特点是模型在训练后保持良性表现，只有在用户发起特定机器遗忘（machine unlearning）请求后才会被激活。现有后门攻击通常在植入后立即生效，容易在利用前被发现；而基于机器遗忘激活的
Formalization of security
该论文是 Gilles Barthe 撰写的一篇关于安全形式化的综述章节，聚焦于证明助手（proof assistants）在计算机安全领域的应用。研究背景是：设计与实现必须满足预期的安全属性，而证明助手能够通过机械化推理严格验证这些属性，
In Audio HAL, there is a possible system becoming unresponsive due to a race condition. This could lead to local denial of service with User execution privileges needed. User in...
CVE-2026-20492
In Audio HAL, there is a possible system becoming unresponsive due to a race condition. This could lead to local denial
In med, there is a possible out of bounds write due to an incorrect bounds check. This could lead to local denial of service with User execution privileges needed. User interact...
CVE-2026-20491
In med, there is a possible out of bounds write due to an incorrect bounds check. This could lead to local denial of ser
In ccci, there is a possible out of bounds read due to a missing bounds check. This could lead to local denial of service if a malicious actor has already obtained the System pr...
CVE-2026-20490
In ccci, there is a possible out of bounds read due to a missing bounds check. This could lead to local denial of servic
In display, there is a possible information disclosure due to an integer overflow. This could lead to local information disclosure if a malicious actor has already obtained the ...
CVE-2026-20489
In display, there is a possible information disclosure due to an integer overflow. This could lead to local information
In display, there is a possible information disclosure due to a missing bounds check. This could lead to local information disclosure if a malicious actor has already obtained t...
CVE-2026-20488
In display, there is a possible information disclosure due to a missing bounds check. This could lead to local informati
In imgsensor, there is a possible application crash due to incorrect error handling. This could lead to local escalation of privilege if a malicious actor has already obtained t...
CVE-2026-20486
In imgsensor, there is a possible application crash due to incorrect error handling. This could lead to local escalation
In HFRP, there is a possible out of bounds write due to a missing bounds check. This could lead to local escalation of privilege if a malicious actor has already obtained the Sy...
CVE-2026-20485
In HFRP, there is a possible out of bounds write due to a missing bounds check. This could lead to local escalation of p
In TFA, there is a possible information disclosure due to a missing permission check. This could lead to local information disclosure if a malicious actor has already obtained t...
CVE-2026-20484
In TFA, there is a possible information disclosure due to a missing permission check. This could lead to local informati
In Telephony, there is a possible escalation of privilege due to a missing permission check. This could lead to local escalation of privilege with no additional execution privil...
CVE-2026-20483
In Telephony, there is a possible escalation of privilege due to a missing permission check. This could lead to local es
In wlan STA FW, there is a possible system becoming unresponsive due to logging. This could lead to remote (proximal/adjacent) denial of service with no additional execution pri...
CVE-2026-20482
In wlan STA FW, there is a possible system becoming unresponsive due to logging. This could lead to remote (proximal/adj
In geniezone, there is a possible out of bounds write due to a missing bounds check. This could lead to local escalation of privilege if a malicious actor has already obtained t...
CVE-2026-20481
In geniezone, there is a possible out of bounds write due to a missing bounds check. This could lead to local escalation
In Audio HAL, there is a possible out of bounds write due to a heap buffer overflow. This could lead to local denial of service with User execution privileges needed. User inter...
CVE-2026-20480
In Audio HAL, there is a possible out of bounds write due to a heap buffer overflow. This could lead to local denial of
In Modem, there is a possible out of bounds read due to a missing bounds check. This could lead to remote denial of service, if a UE has connected to a rogue base station contro...
CVE-2026-20479
In Modem, there is a possible out of bounds read due to a missing bounds check. This could lead to remote denial of serv