English

Relaxed vs. Full Local Decodability with Few Queries: Equivalence and Separations for Linear Codes

Computational Complexity 2025-11-27 v2

Abstract

A locally decodable code (LDC) C ⁣:{0,1}k{0,1}nC \colon \{0,1\}^k \to \{0,1\}^n is an error-correcting code that allows one to recover any bit of the original message with good probability while only reading a small number of bits from a corrupted codeword. A relaxed locally decodable code (RLDC) is a weaker notion where the decoder is additionally allowed to abort and output a special symbol \bot if it detects an error. For a large constant number of queries qq, there is a large gap between the blocklength nn of the best qq-query LDC and the best qq-query RLDC. Existing constructions of RLDCs achieve polynomial length n=k1+O(1/q)n = k^{1 + O(1/q)}, while the best-known qq-LDCs only achieve subexponential length n=2ko(1)n = 2^{k^{o(1)}}. On the other hand, for q=2q = 2, it is known that RLDCs and LDCs are equivalent. We thus ask the question: what is the smallest qq such that there exists a qq-RLDC that is not a qq-LDC? In this work, we show that any linear 33-query RLDC is in fact a 33-LDC, i.e., linear RLDCs and LDCs are equivalent at 33 queries. More generally, we show for any constant qq, there is a soundness error threshold s(q)s(q) such that any linear qq-RLDC with soundness error below this threshold must be a qq-LDC. This implies that linear RLDCs cannot have "strong soundness" -- a stricter condition satisfied by linear LDCs that says the soundness error is proportional to the fraction of errors in the corrupted codeword -- unless they are simply LDCs. In addition, we give simple constructions of linear 1515-query RLDCs that are not qq-LDCs for any constant qq, showing that for q=15q = 15, linear RLDCs and LDCs are not equivalent. We also prove nearly identical results for locally correctable codes and their corresponding relaxed counterpart.

Keywords

Cite

@article{arxiv.2511.02633,
  title  = {Relaxed vs. Full Local Decodability with Few Queries: Equivalence and Separations for Linear Codes},
  author = {Elena Grigorescu and Vinayak M. Kumar and Peter Manohar and Geoffrey Mon},
  journal= {arXiv preprint arXiv:2511.02633},
  year   = {2025}
}
R2 v1 2026-07-01T07:21:23.482Z