Abstract:Let $X$ be a $d$-dimensional simplicial complex. A function $F\colon X(k)\to \{0,1\}^k$ is said to be a direct product function if there exists a function $f\colon X(1)\to \{0,1\}$ such that $F(\sigma) = (f(\sigma_1), \ldots, f(\sigma_k))$ for each $k$-face $\sigma$. In an effort to simplify components of the PCP theorem, Goldreich and Safra introduced the problem of direct product testing, which asks whether one can test if $F\colon X(k)\to \{0,1\}^k$ is correlated with a direct product function by querying $F$ on only $2$ inputs. Dinur and Kaufman conjectured that there exist bounded degree complexes with a direct product test in the small soundness regime. We resolve their conjecture by showing that for all $\delta>0$, there exists a family of high-dimensional expanders with degree $O_{\delta}(1)$ and a $2$-query direct product tester with soundness $\delta$. We use the characterization given by a subset of the authors and independently by Dikstein and Dinur, who showed that some form of non-Abelian coboundary expansion (which they called "Unique-Games coboundary expansion") is a necessary and sufficient condition for a complex to admit such direct product testers. Our main technical contribution is a general technique for showing coboundary expansion of complexes with coefficients in a non-Abelian group. This allows us to prove that the high dimensional expanders constructed by Chapman and Lubotzky satisfies the necessary conditions, thus admitting a 2-query direct product tester with small soundness.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to construct direct product testers with small soundness on high - dimensional expanders (HDX). Specifically, the authors have solved the conjecture proposed by Dinur and Kaufman in 2017, that is, whether there exist high - dimensional expanders with a constant degree that can support 2 - query direct product testers with small soundness. ### Main contributions of the paper 1. **Constructing direct product testers with small soundness**: - The authors prove that for any $\delta>0$ and $\gamma > 0$, there exists a large enough $d$ such that there exists a $d$-dimensional simplicial complex $X$ satisfying the following conditions: - $X$ has $n$ vertices and is a $\gamma$-one - sided local spectral expander. - For $k\geq k_0(\delta)$, the standard direct product tester of $X$ has a soundness parameter $\delta$. - The degree of $X$ does not exceed a certain constant $\Delta$. 2. **Introducing the concept of non - Abelian cohomology expansion**: - The authors define a new non - Abelian cohomology expansion (UG coboundary expansion) and prove that this expansion property is a necessary condition for high - dimensional expanders to support direct product testers with small soundness. - Through this concept, the authors prove that some high - dimensional expanders constructed by Chapman and Lubotzky satisfy the required non - Abelian cohomology expansion property. ### Technical contributions of the paper 1. **Local - to - global cosystolic expansion**: - The authors use a local - to - global method to prove that if all links of a high - dimensional expander are coboundary expanders, then the entire complex itself is also a cosystolic expander. 2. **Vanishing cohomology**: - The authors use the parameter selection of Chapman - Lubotzky complexes to prove that these complexes have vanishing cohomology under certain conditions. 3. **Cohomology expansion of links**: - The authors prove by induction that the links of Chapman - Lubotzky complexes have good cohomology expansion properties. This is one of the main technical contributions of the paper. ### Key techniques and methods of the paper 1. **Cones Method**: - The authors use the Cones Method to construct canonical paths and triangulations, thereby proving that the cohomology constant of a graph has a polynomial relationship with the diameter of the graph. 2. **Extended base cases**: - In addition to handling the base case of $r = 1$, the authors also introduce another base case, that is, when $S_1, S_2, S_3$ satisfy specific conditions, the cohomology constant of the graph $T(S_1, S_2, S_3;\mu)$ can be directly bounded. 3. **$\varepsilon$-product distribution**: - The authors use the concept of $\varepsilon$-product distribution to simplify the problem into a more tractable form, enabling inductive proofs. ### Conclusion This paper successfully constructs direct product testers with small soundness by introducing the new concept of non - Abelian cohomology expansion and combining the local - to - global cosystolic expansion method. This result not only solves the conjecture of Dinur and Kaufman but also provides new tools and methods for other problems in theoretical computer science.

Constant Degree Direct Product Testers with Small Soundness

Characterizing Direct Product Testing via Coboundary Expansion

Quasi-Linear Size PCPs with Small Soundness from HDX

Agreement theorems for high dimensional expanders in the small soundness regime: the role of covers

High-dimensional Expansion of Product Codes is Stronger than Robust and Agreement Testability

High-dimensional expanders from Kac--Moody--Steinberg groups

Low Acceptance Agreement Tests via Bounded-Degree Symplectic HDXs

Coboundary expansion inside Chevalley coset complex HDXs

Coboundary and cosystolic expansion without dependence on dimension or degree

Approaching the Soundness Barrier: A Near Optimal Analysis of the Cube versus Cube Test

Chernoff Bounds and Reverse Hypercontractivity on HDX

Topological expanders, coarse geometry and thick embeddings of complexes

Boolean Function Analysis on High-Dimensional Expanders

New Direct Sum Tests

Lower bounds and integrality gaps in simplicial decomposition

High dimensional expanders and coset geometries

Improved Bounds for High-Dimensional Equivalence and Product Testing using Subcube Queries

From Grassmannian to Simplicial High-Dimensional Expanders

Alternatives for Testing Total Dual Integrality

Coboundary expansion of coset complexes

A Strong Direct Sum Theorem for Distributional Query Complexity