Explicit low-rate construction achieving δ = (1 − ε)/2 and R = Ω(ε^2) without o(1) losses

Construct explicit families of binary linear codes over F2 with relative distance δ = (1 − ε)/2 and rate R ≥ c·ε^2 for some absolute constant c > 0 and all sufficiently small ε > 0, eliminating the current o(1) loss in the best-known explicit constructions.

Background

In the low-rate regime, the GV bound guarantees existence of binary linear codes with relative distance (1−ε)/2 and rate Ω(ε2). Recent explicit constructions achieve δ = (1−ε)/2 with rate Ω(ε{2+o(1)}), but the o(1) term persists. Removing this o(1) term to meet the GV bound exactly in the explicit setting would tighten the best-known constructions to match the existential benchmark.

References

However, there are still open questions. For example, we do not know how to attain δ = \frac{1 - \epsilon}{2} and R = \Omega(\epsilon2) (without any o(1) term) explicitly, and we do not have explicit constructions approaching the GV bound with rates bounded away from zero.

When Do Low-Rate Concatenated Codes Approach The Gilbert-Varshamov Bound? (2405.08584 - Doron et al., 14 May 2024) in Section 1 (Introduction)