Furthermore, we present two pattern sanitization algorithms for blocking inference channels in frequent pattern sharing, and evaluate their performance in the experiments.

• 然后，基于模式净化的思路，提出了两个推理控制算法，并通过实验对算法的性能进行了比较分析。