GV

GeezerVison

The World Wide Web, restored to 800 pixels of glory.

Home|About|Gaming cabinet|Archives|How it works|Starter sites

On this site

  • Algebra
  • Applied Mathematics
  • Calculus and Analysis
  • Discrete Mathematics
  • Foundations of Mathematics
  • Geometry
  • History and Terminology
  • Number Theory
  • Probability and Statistics
  • Recreational Mathematics
  • Topology
  • Alphabetical Index
  • Random Entry
  • New in MathWorld

This view

  • Original page
  • New address

GeezerVison is showing mathworld.wolfram.com — Locally Correctable Code -- from Wolfram MathWorld

Search

Locally Correctable Code


A locally correctable code is an error-correcting code for which a randomized algorithm recovers any requested coordinate of a codeword by reading only a bounded number of coordinates of a corrupted word (Dvir 2016).

For a binary injective encoding C:{0,1}^k->{0,1}^n, a (q,delta,epsilon) locally correctable code has an algorithm that makes at most q queries and, for every message b, requested coordinate u, and word y satisfying d_H(y,C(b))<=deltan, returns C(b)_u with probability at least 1-epsilon. Here d_H is the Hamming distance. Correction recovers a coordinate of the encoded codeword, rather than merely a coordinate of the original message.

Janzer and Manohar (2026) proved that a binary linear code with three-query local correction, fixed delta>0, and fixed epsilon<1/2 must have

 n>=2^(Omega(sqrt(k))),

where Omega is big-omega notation. Reed-Muller codes attain n=2^(O(sqrt(k))), where O is big-O notation, so the dependence on k is optimal up to constant factors in the exponent. The result is specific to binary linear codes and three queries. Janzer and Manohar (2026) credit GPT-6 Astra with completing key parts of their proof approach, which they subsequently simplified.

The definition also extends to codes over infinite fields. For every integer q>=3, put s=|_(q-1)/2_|, where |_x_| is the floor function. Dvir (2026) constructed, for every integer d>=2, a linear code over the real numbers of dimension d and length

 N<=exp(O_q(d^(1/s)lnd)).

It corrects any requested coordinate with probability at least 3/4 using at most q queries when at most N/(8q) coordinates are corrupted. The decoder returns a linear combination of the queried coordinates. The same construction works over every field of characteristic zero. Dvir (2026) attributes the construction to ChatGPT Astra.


See also

Codeword, Error-Correcting Code, Hamming Distance, Linear Code, Reed-Muller Code

Explore with Wolfram|Alpha

WolframAlpha

More things to try:

  • 15.25 + 7.90 + 3.12
  • express 4.8675 through pi and e
  • Laplace transform x^3

References

Dvir, Z. "Constant Query LCCs Over the Reals." 9 Nov 2026. https://arxiv.org/abs/2611.13281v1.Dvir, Z. "Locally Correctable Codes." Lecture 8 in Linear Locally Decodable Codes. 2016. https://www.cs.princeton.edu/~zdvir/LDCnotes/LDC8.pdf.Janzer, O. and Manohar, P. "Reed-Muller Codes Are Optimal Binary Linear 3-Query LCCs." 2 Nov 2026. https://arxiv.org/abs/2611.03145v1.

Cite this as:

Weisstein, Eric W. "Locally Correctable Code." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/LocallyCorrectableCode.html

Subject classifications

You are visitor 01402503 — GeezerVison does not store the pages you look up.
Best viewed with Netscape Navigator 4.7 or Internet Explorer 6 at 800×600. No cookies required.