TL;DR本文研究了关于 Junta 学习问题的变体,分析了广义情况及量子查询复杂度,并展示了当预定义函数为 OR 函数或精确一半函数时最优的解决方案,对于大多数函数实现了二次优化。
Abstract
In this paper, we study the following variant of the junta learning problem.
We are given oracle access to a Boolean function $f$ on $n$ variables that only
depends on $k$ variables, and, when restricted to them, equals some predefined
function $h$. The task is to identify the variable