We study testing $\pi$-freeness of functions $f:[n]^d\to\mathbb{R}$, where $f$ is $\pi$-free if there there are no $k$ indices $x_1\prec\cdots\prec x_k\in [n]^d$ such that $f(x_i)<f(x_j)$ and $\pi(i) < \pi(j)$ for all $i,j \in [k]$, where $\prec$ is the natural partial order over $[n]^d$. Given $\epsilon\in(0,1)$, $\epsilon$-testing $\pi$-freeness asks to distinguish $\pi$-free functions from those which are $\epsilon$-far -- meaning at least $\epsilon n^d$ function values must be modified to make it $\pi$-free. While $k=2$ coincides with monotonicity testing, far less is known for $k>2$. We initiate a systematic study of pattern freeness on higher-dimensional grids. For $d=2$ and all permutations of size $k=3$, we design an adaptive one-sided tester with query complexity $O(n^{4/5+o(1)})$. We also prove general lower bounds for $k=3$: every nonadaptive tester requires $\Omega(n)$ queries, and every adaptive tester requires $\Omega(\sqrt{n})$ queries, yielding the first super-logarithmic lower bounds for $\pi$-freeness. For the monotone patterns $\pi=(1,2,3)$ and $(3,2,1)$, we present a nonadaptive tester with polylogarithmic query complexity, giving an exponential separation between monotone and nonmonotone patterns (unlike the one-dimensional case). A key ingredient in our $\pi$-freeness testers is new erasure-resilient ($\delta$-ER) $\epsilon$-testers for monotonicity over $[n]^d$ with query complexity $O(\log^{O(d)}n/(\epsilon(1-\delta)))$, where $0<\delta<1$ is an upper bound on the fraction of erasures. Prior ER testers worked only for $\delta=O(\epsilon/d)$. Our nonadaptive monotonicity tester is nearly optimal via a matching lower bound due to Pallavoor, Raskhodnikova, and Waingarten (Random Struct. Algorithms, 2022). Finally, we show that current techniques cannot yield sublinear-query testers for patterns of length $4$ even on two-dimensional hypergrids.
翻译:我们研究函数 $f:[n]^d\to\mathbb{R}$ 的 $\pi$-无性测试,其中若不存在 $k$ 个索引 $x_1\prec\cdots\prec x_k\in [n]^d$ 使得对于所有 $i,j \in [k]$ 有 $f(x_i)<f(x_j)$ 且 $\pi(i) < \pi(j)$,则称 $f$ 是 $\pi$-无的,这里 $\prec$ 是 $[n]^d$ 上的自然偏序。给定 $\epsilon\in(0,1)$,$\epsilon$-测试 $\pi$-无性要求区分 $\pi$-无函数与那些 $\epsilon$-远的函数——后者意味着至少需要修改 $\epsilon n^d$ 个函数值才能使其成为 $\pi$-无的。当 $k=2$ 时,该问题等同于单调性测试,但对于 $k>2$ 的情况所知甚少。我们开启了高维网格上模式无性研究的系统工作。对于 $d=2$ 且所有大小为 $k=3$ 的排列,我们设计了一个查询复杂度为 $O(n^{4/5+o(1)})$ 的自适应单边测试器。我们还证明了 $k=3$ 时的一般下界:每个非自适应测试器需要 $\Omega(n)$ 次查询,每个自适应测试器需要 $\Omega(\sqrt{n})$ 次查询,这给出了 $\pi$-无性的首个超对数下界。对于单调模式 $\pi=(1,2,3)$ 和 $(3,2,1)$,我们提出了一个具有多对数查询复杂度的非自适应测试器,从而在单调模式与非单调模式之间建立了指数级分离(与一维情况不同)。我们 $\pi$-无性测试器的一个关键组成部分是针对 $[n]^d$ 上单调性的新型擦除弹性($\delta$-ER)$\epsilon$-测试器,其查询复杂度为 $O(\log^{O(d)}n/(\epsilon(1-\delta)))$,其中 $0<\delta<1$ 是擦除比例的上界。先前的 ER 测试器仅适用于 $\delta=O(\epsilon/d)$。我们的非自适应单调性测试器通过 Pallavoor、Raskhodnikova 和 Waingarten(Random Struct. Algorithms, 2022)给出的匹配下界证明是近乎最优的。最后,我们表明即使对于二维超网格,现有技术也无法为长度为 $4$ 的模式产生亚线性查询测试器。