LIPIcs.ITCS.2025.82.pdf
- Filesize: 0.85 MB
- 21 pages
In this work, we consider the list-decodability and list-recoverability of codes in the zero-rate regime. Briefly, a code π β [q]βΏ is (p,π,L)-list-recoverable if for all tuples of input lists (Yβ,β¦ ,Y_n) with each Y_i β [q] and |Y_i| = π, the number of codewords c β π such that c_i β Y_i for at most pn choices of i β [n] is less than L; list-decoding is the special case of π = 1. In recent work by Resch, Yuan and Zhang (ICALP 2023) the zero-rate threshold for list-recovery was determined for all parameters: that is, the work explicitly computes p_*: = p_*(q,π,L) with the property that for all Ξ΅ > 0 (a) there exist positive-rate (p_*-Ξ΅,π,L)-list-recoverable codes, and (b) any (p_*+Ξ΅,π,L)-list-recoverable code has rate 0. In fact, in the latter case the code has constant size, independent on n. However, the constant size in their work is quite large in 1/Ξ΅, at least |π| β₯ (1/(Ξ΅))^O(q^L). Our contribution in this work is to show that for all choices of q,π and L with q β₯ 3, any (p_*+Ξ΅,π,L)-list-recoverable code must have size O_{q,π,L}(1/Ξ΅), and furthermore this upper bound is complemented by a matching lower bound Ξ©_{q,π,L}(1/Ξ΅). This greatly generalizes work by Alon, Bukh and Polyanskiy (IEEE Trans. Inf. Theory 2018) which focused only on the case of binary alphabet (and thus necessarily only list-decoding). We remark that we can in fact recover the same result for q = 2 and even L, as obtained by Alon, Bukh and Polyanskiy: we thus strictly generalize their work. Our main technical contribution is to (a) properly define a linear programming relaxation of the list-recovery condition over large alphabets; and (b) to demonstrate that a certain function defined on a q-ary probability simplex is maximized by the uniform distribution. This represents the core challenge in generalizing to larger q (as a binary simplex can be naturally identified with a one-dimensional interval). We can subsequently re-utilize certain Schur convexity and convexity properties established for a related function by Resch, Yuan and Zhang along with ideas of Alon, Bukh and Polyanskiy.
Feedback for Dagstuhl Publishing