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

Degree lower bounds without hyper-regularity

Develop degree lower bounds for skeletons of high dimensional expanders that remove the hyper-regularity assumption. Specifically, prove super-exponential degree lower bounds for families such as Ramanujan complexes whose links are extremely unbalanced, or otherwise identify refined concentration notions that enable extending the current technique beyond the hyper-regular case.

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

Background

The paper establishes super-exponential degree lower bounds for certain hyper-regular HDX using sampling results. However, many prominent constructions, including Ramanujan complexes, have highly unbalanced links, and the current method does not apply.

The authors suspect this limitation is technical and suggest that sharper concentration tools might remove the hyper-regularity constraint.

References

The best known constructions of high dimensional expanders are not hyper-regular. Is it possible to remove this constraint from our degree lower bound? While our technique holds even for `reasonably balanced' complexes, it cannot handle objects like the Ramanujan complexes have extremely unbalanced links. It seems likely this is a technical rather than inherent barrier, and we conjecture some finer notion of concentration or application thereof may be able to remove this constraint.

Chernoff Bounds and Reverse Hypercontractivity on HDX (2404.10961 - Dikstein et al., 17 Apr 2024) in Open questions section