跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2604.05615 · v2

Classes Testable with $O(1/ε)$ Queries for $ε$ Below an $n$-Independent Threshold

首次提交 2026-04-07 17:04
版本更新 2026-11-12 18:04
雷达观测 2026-11-13 14:15
arXiv ↗PDF ↗

摘要

In this paper, we study classes of Boolean functions that are testable with O(ψ+1/ε)O(ψ+1/ε) queries, where ψψ depends on the parameters of the class (e.g., the number of terms, the number of relevant variables, etc.), is independent of (or sublinear in) 1/ε1/ε, and is independent of the total number of variables nn. In particular, when εε is sufficiently small so that ε=O(1/ψ)ε=O(1/ψ), the query complexity is O(1/ε)O(1/ε), matching the known tight lower bound Ω(1/ε)Ω(1/ε). In particular, if the structural parameters of the class are fixed and ε=1/ωn(1)ε=1/ω_n(1) (e.g., 1/log⁡n1/\log n), then the asymptotic query complexity is O(1/ε)O(1/ε). This result was previously known for classes of terms of size at most kk and exclusive OR functions of at most kk variables. In this paper, we extend this list to include the classes: kk-junta, functions with Fourier degree at most dd, ss-sparse polynomials of degree at most dd, and ss-sparse polynomials. Additionally, we show that for any class CC of Boolean functions that depend on at most kk variables, if CC is properly exactly learnable, then it is testable with O(1/ε)O(1/ε) queries for ε<1/ψε<1/ψ, where ψψ depends on kk and independent of the total number of variables nn.

作者备注:Appears in the Proceedings of the 52nd International Conference on Current Trends in Theory and Practice of Computer Science (Sofsem 2027)

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-12 18:04 · 最新观测版本

Classes Testable with $O(1/ε)$ Queries for $ε$ Below an $n$-Independent Threshold

In this paper, we study classes of Boolean functions that are testable with O(ψ+1/ε)O(ψ+1/ε) queries, where ψψ depends on the parameters of the class (e.g., the number of terms, the number of relevant variables, etc.), is independent of (or sublinear in) 1/ε1/ε, and is independent of the total number of variables nn. In particular, when εε is sufficiently small so that ε=O(1/ψ)ε=O(1/ψ), the query complexity is O(1/ε)O(1/ε), matching the known tight lower bound Ω(1/ε)Ω(1/ε). In particular, if the structural parameters of the class are fixed and ε=1/ωn(1)ε=1/ω_n(1) (e.g., 1/log⁡n1/\log n), then the asymptotic query complexity is O(1/ε)O(1/ε). This result was previously known for classes of terms of size at most kk and exclusive OR functions of at most kk variables. In this paper, we extend this list to include the classes: kk-junta, functions with Fourier degree at most dd, ss-sparse polynomials of degree at most dd, and ss-sparse polynomials. Additionally, we show that for any class CC of Boolean functions that depend on at most kk variables, if CC is properly exactly learnable, then it is testable with O(1/ε)O(1/ε) queries for ε<1/ψε<1/ψ, where ψψ depends on kk and independent of the total number of variables nn.

查看此版本解读 →查看此版本原文 ↗