Abstract:We prove a general structural theorem for a wide family of local algorithms, which includes property testers, local decoders, and PCPs of proximity. Namely, we show that the structure of every algorithm that makes $q$ adaptive queries and satisfies a natural robustness condition admits a sample-based algorithm with $n^{1- 1/O(q^2 \log^2 q)}$ sample complexity, following the definition of Goldreich and Ron (TOCT 2016). We prove that this transformation is nearly optimal. Our theorem also admits a scheme for constructing privacy-preserving local algorithms. Using the unified view that our structural theorem provides, we obtain results regarding various types of local algorithms, including the following. - We strengthen the state-of-the-art lower bound for relaxed locally decodable codes, obtaining an exponential improvement on the dependency in query complexity; this resolves an open problem raised by Gur and Lachish (SICOMP 2021). - We show that any (constant-query) testable property admits a sample-based tester with sublinear sample complexity; this resolves a problem left open in a work of Fischer, Lachish, and Vasudev (FOCS 2015) by extending their main result to adaptive testers. - We prove that the known separation between proofs of proximity and testers is essentially maximal; this resolves a problem left open by Gur and Rothblum (ECCC 2013, Computational Complexity 2018) regarding sublinear-time delegation of computation. Our techniques strongly rely on relaxed sunflower lemmas and the Hajnal-Szemerédi theorem.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the general theorem regarding the structure of local algorithms, especially for algorithms with natural robustness conditions, which include property testers, local decoders, and probabilistically checkable proofs (PCPs) of proximity. Specifically, the main objectives of the paper can be summarized as follows: 1. **Prove the general structure theorem**: The paper proves that for any local algorithm that makes $q$ adaptive queries and satisfies natural robustness conditions, there exists a sample - based algorithm with a sample complexity of $n^{1 - 1/O(q^2\log^2q)}$. This result shows that through appropriate transformation, many local algorithms can be converted into sample - based algorithms, thus simplifying the design and analysis of algorithms in some cases. 2. **Demonstrate the approximate optimality of the transformation**: The paper also proves that the above - mentioned transformation is almost optimal, that is, any such transformation must generate an algorithm with a sample complexity of at least $n^{1 - 1/\Omega(q)}$. This indicates that the transformation method proposed in the paper is theoretically very close to the optimal. 3. **Construct privacy - protected local algorithms**: The paper proposes a scheme for constructing privacy - protected local algorithms. These algorithms can perform calculations without revealing the specific information of the calculated function, which is very important for privacy - protected application scenarios. 4. **Apply to different types of local algorithms**: Using the above - mentioned structure theorem, the paper obtains several important results in the fields of coding theory, property testing, and verification: - **Enhance the lower bound of local decoding codes**: The paper significantly improves the query complexity dependence of relaxed locally decodable codes (RLDCs), solving an open problem proposed by Gur and Lachish. - **Property testing with sub - linear sample complexity**: The paper proves that any testable property with a constant number of query times can be tested by a sample - based tester with a sample complexity of $n^{1 - 1/O(q^2\log^2q)}$, solving an open problem left by Fischer, Lachish, and Vasudev. - **Maximum separation between verification and testing**: The paper proves that the known separation between proof - of - proximity systems and testers is essentially the maximum, solving an open problem proposed by Gur and Rothblum. In conclusion, by establishing a general structure theorem, this paper not only theoretically advances the understanding of local algorithms but also provides important technical tools and methods in multiple practical application scenarios.

A Structural Theorem for Local Algorithms with Applications to Coding, Testing, and Verification

Local Verifications in Distributed Data Structures

High rate locally-correctable and locally-testable codes with sub-polynomial query complexity

Relaxed Local Correctability from Local Testing

New Techniques and Tighter Bounds for Local Computation Algorithms

Simple Local Computation Algorithms for the General Lovasz Local Lemma

An Algorithmic Proof of the Lovasz Local Lemma via Resampling Oracles

An Exponential Lower Bound for Linear 3-Query Locally Correctable Codes

Local Decoding in Distributed Approximate Computing

On Good $2$-Query Locally Testable Codes from Sheaves on High Dimensional Expanders

On Local Distributed Sampling and Counting

Locally Testable Codes and Cayley Graphs.

Lower bounds for constant query affine-invariant LCCs and LTCs

Distributed Approximate Computing with Constant Locality

Query complexity lower bounds for local list-decoding and hard-core predicates (even for small rate and huge lists)

The Topology of Local Computing in Networks

Geometrically Local Quantum and Classical Codes from Subdivision

What Can Be Sampled Locally?

Local certification of local properties: tight bounds, trade-offs and new parameters

Local Proofs Approaching the Witness Length

On Derandomizing Local Distributed Algorithms