[Paper Review] Explicit rank-metric codes list-decodable with optimal redundancy

Venkatesan Guruswami, Carol Wang|arXiv (Cornell University)|Nov 27, 2013

Coding theory and cryptography20 references7 citations

TL;DR

This paper presents the first explicit construction of linear rank-metric codes that are list-decodable up to a fraction $1 - R - \varepsilon$ of rank errors, where $R$ is the code rate and $\varepsilon > 0$. By restricting message polynomials to a carefully chosen $\mathbb{F}_h$-subspace that evades structured subspaces from the linear-algebraic list decoder, the authors achieve optimal redundancy with a polynomial list size, leveraging subspace designs and subspace-evasive sets.

ABSTRACT

We construct an explicit family of linear rank-metric codes over any field ${\mathbb F}_h$ that enables efficient list decoding up to a fraction $ρ$ of errors in the rank metric with a rate of $1-ρ-ε$, for any desired $ρ\in (0,1)$ and $ε> 0$. Previously, a Monte Carlo construction of such codes was known, but this is in fact the first explicit construction of positive rate rank-metric codes for list decoding beyond the unique decoding radius. Our codes are subcodes of the well-known Gabidulin codes, which encode linearized polynomials of low degree via their values at a collection of linearly independent points. The subcode is picked by restricting the message polynomials to an ${\mathbb F}_h$-subspace that evades the structured subspaces over an extension field ${\mathbb F}_{h^t}$ that arise in the linear-algebraic list decoder for Gabidulin codes due to Guruswami and Xing (STOC'13). This subspace is obtained by combining subspace designs contructed by Guruswami and Kopparty (FOCS'13) with subspace evasive varieties due to Dvir and Lovett (STOC'12). We establish a similar result for subspace codes, which are a collection of subspaces, every pair of which have low-dimensional intersection, and which have received much attention recently in the context of network coding. We also give explicit subcodes of folded Reed-Solomon (RS) codes with small folding order that are list-decodable (in the Hamming metric) with optimal redundancy, motivated by the fact that list decoding RS codes reduces to list decoding such folded RS codes. However, as we only list decode a subcode of these codes, the Johnson radius continues to be the best known error fraction for list decoding RS codes.

Motivation & Objective

• To construct explicit, high-rate rank-metric codes list-decodable beyond the unique decoding radius.
• To overcome the challenge of large field sizes in Gabidulin codes by using subspace-evasive subcodes.
• To extend the construction to subspace codes and folded Reed-Solomon codes for Hamming metric list decoding.
• To provide an explicit alternative to prior Monte Carlo constructions of list-decodable codes with optimal redundancy.
• To establish the first explicit construction of positive-rate rank-metric codes list-decodable from more than half the minimum distance.

Proposed method

• Construct explicit subcodes of Gabidulin codes by restricting message polynomials to an $\mathbb{F}_h$-subspace that avoids structured subspaces from the linear-algebraic list decoder.
• Use subspace designs from Guruswami and Kopparty (FOCS'13) to control intersections with candidate list decoder outputs.
• Incorporate subspace-evasive varieties from Dvir and Lovett (STOC'12) to ensure the subcode has small intersection with the decoder's output subspace.
• Apply the construction to folded Reed-Solomon codes with low folding order to achieve list-decoding with optimal redundancy in the Hamming metric.
• Leverage the fact that list-decoding folded RS codes reduces to list-decoding subcodes with small folding order.
• Ensure the final list is contained in an $\mathbb{F}_h$-subspace of dimension $O(s^2/\varepsilon^2)$, where $s$ controls the error fraction and $\varepsilon$ controls rate loss.

Experimental results

Research questions

• RQ1 Can explicit, high-rate rank-metric codes be constructed that are list-decodable beyond the unique decoding radius?
• RQ2 Is it possible to achieve optimal redundancy in list-decoding rank-metric codes using explicit constructions rather than Monte Carlo methods?
• RQ3 Can the same techniques be extended to subspace codes and folded Reed-Solomon codes for Hamming metric list decoding?
• RQ4 What is the minimal list size achievable for explicit list-decodable rank-metric codes with redundancy $1 - R - \varepsilon$?
• RQ5 Can subspace-evasive sets be used to construct codes with poly(1/ε) list size for error fraction $1 - R - \varepsilon$?

Key findings

• The paper constructs an explicit $\mathbb{F}_h$-linear subcode of a Gabidulin code with rate $(1 - \varepsilon)k/n$ and list-decoding radius $s(n - k)/(s + 1)$ rank errors.
• The final list of codewords is contained in an $\mathbb{F}_h$-subspace of dimension $O(s^2/\varepsilon^2)$, ensuring polynomial list size.
• The construction achieves a list-decoding radius approaching the information-theoretic limit of $1 - R$ as $s$ increases.
• For subspace codes, the paper gives the first explicit construction of high-rate, list-decodable codes beyond the unique decoding radius, without using folding.
• The method applies to folded Reed-Solomon codes with low folding order, yielding a subcode list-decodable up to list-decoding capacity with optimal redundancy.
• The construction improves upon prior explicit codes that only achieved polynomially small rates, and provides a deterministic alternative to Monte Carlo constructions.