Dice Question Streamline Icon: https://streamlinehq.com

Polynomially small redundancy for small-value augmented retrieval

Determine whether augmented retrieval with value domain size V = n^{o(1)} (including the case where V is a power of two) admits constant-time queries with polynomially small redundancy n^{1−Ω(1)}.

Information Square Streamline Icon: https://streamlinehq.com

Background

The paper introduces augmented retrieval, showing that when an array of augmented queries is stored alongside retrieval, the redundancy can be reduced dramatically for large value universes (e.g., V ≥ n{3+O(1)}), with constant-time queries and near-optimal space.

For smaller V, the provided constructions either require larger redundancy or additional randomness assumptions. The authors explicitly pose whether polynomially small redundancy remains achievable when V is subpolynomial in n, which would require new algorithmic ideas beyond those developed in the paper.

References

Whether or not one can achieve polynomially small redundancy $n{1 - \Omega(1)}$ for $V = n{o(1)}$ (even assuming, for example, that $V$ is a power of two) appears to remain as an open question, and appears to be a problem that will require significant additional algorithmic ideas to solve.

Static Retrieval Revisited: To Optimality and Beyond (2510.18237 - Hu et al., 21 Oct 2025) in Section 6 (Open Problems)