Table of contents 介绍 前言 一、差分隐私的承诺 隐私保护的数据分析 参考文献 二、基本术语 计算模型 定义隐私数据分析 形式化差分隐私 形式化差分隐私(1) 形式化差分隐私(2) 差分隐私定义补充说明 参考文献 三、基本技术与合成定理 概率工具 随机响应 Laplace机制 Exponential机制 合成定理 3.5.1 合成:一些技术细节 3.5.2 高级合成技术 3.5.3 拉普拉斯vs高斯 3.5.4 合成定理注释 稀疏向量技术 3.6.1 高于阈值算法 3.6.2 稀疏算法 3.6.3 数值稀疏算法 参考文献 四、带相关误差的线性查询 SmallDB:离线算法 SmallDB:精确边界 可乘权重算法:在线机制 可乘权重算法更新规则 数值稀疏向量可乘权重算法 参考文献 五、差分隐私泛化 𝞪-nets机制 迭代构建机制 应用:其他数据库更新算法 迭代构建机制和在线算法 机制关联 迭代构建机制和𝞪-nets 不可知论学习 查询发布的博弈论观点 参考文献 六、查询增强 查询算法的Boosting 七、非典型最坏敏感度 八、差分隐私下界与分离结果 九、差分隐私与计算复杂性 十、差分隐私与机制设计 十一、差分隐私与机器学习 十二、其他差分隐私模型 十三、思考