论文标题
建立差异关系隐私及其在有关回答中的使用
Towards Differential Relational Privacy and its use in Question Answering
论文作者
论文摘要
在使用训练有素的模型来回答时，数据集中实体之间关系的记忆可能会导致隐私问题。我们介绍了关系记忆（RM），以理解，量化和控制这一现象。虽然界限一般的记忆可能会对训练有素的模型的性能产生不利影响，但边界RM并不能阻止有效学习。当数据分布长尾时，差异最为明显，许多查询只有很少的培训示例：阻碍一般的记忆阻止有效学习，同时仅阻碍关系记忆仍然可以学习基础概念的一般属性。我们正式化了关系隐私的概念（RP），并受到差异隐私（DP）的启发，我们提供了差异关系隐私（DRP）的可能定义。这些概念可用于描述和计算受过训练的模型中RM数量的界限。我们说明了与大规模模型的实验中的关系隐私概念，以回答问题。
Memorization of the relation between entities in a dataset can lead to privacy issues when using a trained model for question answering. We introduce Relational Memorization (RM) to understand, quantify and control this phenomenon. While bounding general memorization can have detrimental effects on the performance of a trained model, bounding RM does not prevent effective learning. The difference is most pronounced when the data distribution is long-tailed, with many queries having only few training examples: Impeding general memorization prevents effective learning, while impeding only relational memorization still allows learning general properties of the underlying concepts. We formalize the notion of Relational Privacy (RP) and, inspired by Differential Privacy (DP), we provide a possible definition of Differential Relational Privacy (DrP). These notions can be used to describe and compute bounds on the amount of RM in a trained model. We illustrate Relational Privacy concepts in experiments with large-scale models for Question Answering.
