Manifest-Assisted Retrieval in Artificial Intelligence Search: Evidence Gaps and a Security Evaluation
The external sources retained in our non-systematic audit did not isolate the incremental security or utility effect of
#cs.DB
共收录 43 条相关情报。
← 返回所有主题Revision-Bound Evidence Cuts: Authenticated Queries and Contextual Witness Capacity
Authentic records can support unauthorized narratives when their qualifications, query scope, or counterevidence are sil
Constraining the Commit Boundary: A DIO/ODI-Informed Authorization Guardrail for Agentic Entity Resolution
Agentic Entity Resolution (ER) systems iteratively plan actions, retrieve evidence, revise hypotheses, and propose graph
Decentralized Medical Data Sharing with Off-chain Storage and View-based Access Control
Recently, many blockchain-based medical data sharing systems using off-chain storage have been proposed. However, these
HNSecW: A Logarithmic-Time MPC Protocol for Approximate Nearest Neighbor Search
As a core component of modern RAG systems, approximate nearest neighbor (ANN) search has been extensively studied for de
Holo-NN: Holographic Nearest Neighbor Search via Compressive Sensing
In the age of AI, privacy-preserving approximate nearest neighbor (ANN) search over high-dimensional embeddings has emer
Amora: Communication-Efficient Private Approximate Nearest Neighbors Search
Approximate nearest neighbors search (ANNS) over cloud-hosted vector databases raises a critical privacy concern: the qu
ARACDE: Retrofitting Databases with ABAC using External Authorization Engines
This paper proposes a methodology for decoupling the authorization logic of a relational database from its core data man
Risk-aware Road-network Trajectory Synthesis under Local Differential Privacy
Road-network trajectory synthesis under local differential privacy (LDP) enables mobility data analysis without uploadin
Soft Voting for Policy-Aware Private Data Synthesis
Blowfish privacy relaxes differential privacy (DP) by protecting only the attribute-value substitutions a data owner spe
Bounded Provisional Visibility: Controlling Poisoning Exposure in Continuously Ingested RAG Vector Stores
Continuous ingestion can expose new retrieval-augmented generation (RAG) content to retrieval before vetting completes,
Epistemic Typing as a PostgreSQL Table Access Method: Adversarial Conflict Resolution Under Confidence Forgery and Sybil Coordination
该论文提出了 KNDB——一个以 PostgreSQL 18 表访问方法（Table Access Method, TAM）形式实现的「认知类型化」引擎，用于在数据库写入路径上直接完成对抗性冲突消解。研究背景是数据融合与众包真值发现（tru
Agentic Network Traffic Monitoring
随着 agentic AI 在各行业广泛应用，攻击面不断扩大，确保 AI agent 的行为符合用户意图成为关键安全需求。审计 agent 的网络流量可以提供清晰的交互记录。本文提出一种基于复值超稀疏流量矩阵（complex valued
Subjects, Not Authors: The Authorship Hazard in Agentic Dataspaces
这篇论文研究的是数据空间（dataspace）与 LLM 智能体交汇处的一个治理盲区。作者指出，数据空间连接器（connector）的决策点只决定「是否允许一次传输发生」，而不管被传输的值里到底装了什么。对基于合同的传统应用来说这尚可接受，
Pattern-level Differential Privacy for High-utility Complex Event Processing
该论文聚焦复杂事件处理（CEP）系统中的隐私保护问题。CEP 引擎从高速事件流中检测模式（如事件序列、时序关联、聚合条件），而现有的隐私保护机制（PPM）通常对整条数据流或单个事件施加统一噪声，限制性过强：隐私预算被平摊到大量与最终模式无关
A Policy Profile for Croissant: Refusal as a Property of the Dataset
Croissant 是机器学习数据集领域事实上的机器可读描述符标准，其形式为基于 schema.org 的 JSON-LD。自 1.1 版本起，Croissant 开始承载数据使用条件（data use conditions），并推荐使用
The Stochastic Deputy: Structural Tenant Isolation for Tool-Using LLM Agents
该论文形式化了工具调用型 LLM Agent 中的「随机代理问题」(stochastic deputy problem):多租户工具通常要求调用方传入租户标识(tenant id),再由工具依据调用方授权校验该标识。但当调用方是 LLM A
You've Got a BUD in Me: Authenticated Reads from Per-Block Write Logs
该论文研究的是区块链「认证读取」(authenticated reads) 的可扩展性问题。传统做法是维护一棵覆盖全量状态的认证数据结构(如 Merkle Patricia Trie，MPT)，以此支撑历史成员性证明与排除性证明；但随着状态
Poisoning Attacks on the PGM-index
本论文研究了一种针对学习型索引结构 PGM-index 的投毒攻击方法。PGM-index 是一种基于最优分段线性近似（PLA）的实用学习索引，通过最小化分段数量来提升查询效率。作者首次探讨了这种最优 PLA 本身对恶意数据插入的敏感性，提
Local Private Information Retrieval for Graph-Based Replicated Systems
本文重新审视了多服务器图复制系统中的私有信息检索(PIR)隐私定义，提出一种仅由服务器存储结构决定的、更贴合实际需求的“局部用户隐私”(local user privacy)模型。经典图复制PIR要求用户检索某条消息时，必须对所有服务器隐藏