Abstract:The multiplicity Schwartz-Zippel lemma asserts that over a field, a low-degree polynomial cannot vanish with high multiplicity very often on a sufficiently large product set. Since its discovery in a work of Dvir, Kopparty, Saraf and Sudan [SIAM J. Comput., 2013], the lemma has found numerous applications in both math and computer science; in particular, in the definition and properties of multiplicity codes by Kopparty, Saraf and Yekhanin [J. ACM, 2014]. In this work, we show how to algorithmize the multiplicity Schwartz-Zippel lemma for arbitrary product sets over any field. In other words, we give an efficient algorithm for unique decoding of multivariate multiplicity codes from half their minimum distance on arbitrary product sets over all fields. Previously, such an algorithm was known either when the underlying product set had a nice algebraic structure: for instance, was a subfield (by Kopparty [ToC, 2015]) or when the underlying field had large (or zero) characteristic, the multiplicity parameter was sufficiently large and the multiplicity code had distance bounded away from $1$ (Bhandari, Harsha, Kumar and Sudan [STOC 2021]). In particular, even unique decoding of bivariate multiplicity codes with multiplicity two from half their minimum distance was not known over arbitrary product sets over any field. Our algorithm builds upon a result of Kim and Kopparty [ToC, 2017] who gave an algorithmic version of the Schwartz-Zippel lemma (without multiplicities) or equivalently, an efficient algorithm for unique decoding of Reed-Muller codes over arbitrary product sets. We introduce a refined notion of distance based on the multiplicity Schwartz-Zippel lemma and design a unique decoding algorithm for this distance measure. On the way, we give an alternate analysis of Forney's classical generalized minimum distance decoder that might be of independent interest.

What problem does this paper attempt to address?

The core problem that this paper attempts to solve is the algorithmic Multiplicity Schwartz - Zippel Lemma, that is, to provide an efficient unique decoding algorithm for multivariate multiplicity codes on any product set. Specifically, this problem can be divided into the following aspects: 1. **Background and Motivation**: - **Schwartz - Zippel Lemma**: This is a fundamental conclusion that a low - degree polynomial cannot vanish frequently with high multiplicity on a sufficiently large product set. - **Multiplicity Schwartz - Zippel Lemma**: This is a generalization of the Schwartz - Zippel Lemma, taking into account the derivative information (i.e., multiplicity) of the polynomial. This lemma shows that on a sufficiently large product set, a low - degree polynomial cannot vanish frequently with high multiplicity. 2. **Limitations of Existing Work**: - Previous work has mainly focused on product sets with special algebraic structures (such as sub - fields) or under specific conditions (such as large - characteristic fields, sufficiently large multiplicity parameters, etc.). - Even for bivariate multiplicity codes and the case where the multiplicity parameter is 2, there is a lack of effective unique decoding algorithms. 3. **Main Contributions of the Paper**: - **Algorithmic Multiplicity Schwartz - Zippel Lemma**: The author proposes an efficient algorithm applicable to any product set and any field for uniquely decoding multivariate multiplicity codes. - **New Distance Definition**: A refined notion of distance based on multiplicity is introduced, and the corresponding unique decoding algorithm is designed. - **Alternative Analysis of Forney GMD Decoding**: An alternative analysis of Forney's Generalized Minimum Distance Decoding is provided, which may be of independent interest. 4. **Specific Problem Description**: - Given a received word $ f : T^m \to F_{<s}[z_1, \ldots, z_m] $, it is required to find the unique polynomial $ P \in F[x_1, \ldots, x_m] $ (if it exists) such that the distance between $ \text{Enc}^{(s)}(P) $ and $ f $ on $ T^m $ is less than the proportion of $\frac{1}{2}\left(1 - \frac{d}{s|T|}\right)$. - Here, $ \text{Enc}^{(s)}(P)(a) := P(a + z)\ mod\ \langle z \rangle^s $, which represents the values of the polynomial $ P $ at point $ a $ and all of its partial derivatives of order not exceeding $ s - 1$. 5. **Algorithm Overview**: - The algorithm is based on the generalization of the Reed - Muller code decoder by Kim and Kopparty, but it needs to deal with more complex multiplicity problems. - A new distance metric is introduced, and a weighted univariate multiplicity code decoder is used to extract the polynomial $ P_\ell $. Through these improvements, this paper solves the problem of uniquely decoding multivariate multiplicity codes on any product set and any field, fills the gaps in existing methods, and provides new tools and perspectives for further research.

Algorithmizing the Multiplicity Schwartz-Zippel Lemma

Polynomials, Divided Differences, and Codes

Fast list-decoding of univariate multiplicity and folded Reed-Solomon codes

Feasibly Constructive Proof of Schwartz-Zippel Lemma and the Complexity of Finding Hitting Sets

Syndrome decoding of Reed-Muller codes and tensor decomposition over finite fields

List Decoding Tensor Products and Interleaved Codes

Lifted Reed-Solomon Codes and Lifted Multiplicity Codes

List Decoding Reed-Solomon, Algebraic-Geometric, and Gabidulin Subcodes Up to the Singleton Bound.

High Rate Multivariate Polynomial Evaluation Codes

Syndrome decoding meets multiple instances

Decodable quantum LDPC codes beyond the square root distance barrier using high dimensional expanders

Maximum Sum-Rank Distance Codes over Finite Chain Rings

Quantum Algorithms for Representation-Theoretic Multiplicities

Efficient Multi-Point Local Decoding of Reed-Muller Codes Via Interleaved Codex

List Decoding of Tanner and Expander Amplified Codes from Distance Certificates

AG Codes Achieve List-decoding Capacity over Constant-sized Fields

Information Sets of Multiplicity Codes

On Multi-Point Local Decoding of Reed-Muller Codes.

Distance Distribution in Reed-Solomon Codes

Randomized Decoding of Linearized Reed-Solomon Codes Beyond the Unique Decoding Radius