n^{5/2} threshold for 3-graph homeomorphs
Conjecture 1.2 · arXiv:2004.02657
Status partial high confidence
Conjecture 1.2 posits that the optimal exponent for homeomorphs of any fixed 3-graph $H$ is $5/2$; the source paper itself establishes the weaker universal bound $3-1/5 \approx 2.8$. Maya Sankar (2024) subsequently improved this upper bound to $8/3 \approx 2.667$ for all 2-dimensional simplicial complexes, bringing it closer to the conjectured optimum. The full conjecture remains open.
Cited literature (1)
-
Proves that any $n$-vertex 2-dimensional simplicial complex with at least $C_S n^{8/3}$ facets contains a homeomorphic copy of any fixed 2-complex $S$, improving the universal exponent from $3-1/5$ toward the conjectured optimal $5/2$.
Reviewer notes. The Long-Narayanan-Yap paper (Discrete Analysis, ~2020) extends the universal exponent framework to all dimensions k with exponent $\lambda_k \geq k^{-2k^2}$, but does not improve the 2-dimensional case over the source paper. Sankar (2024) is the primary improvement for 2-complexes specifically. The conjecture that $5/2$ is the correct universal exponent for 3-graph homeomorphs remains open.
Context
The paper proves the exponent $3-1/5$ is universal but does not believe it to be tight. The conjectured optimal exponent $5/2 = 3 - 1/2$ is motivated by the Brown–Erdős–Sós result that $5/2$ is the correct exponent for the sphere $S^2$ (i.e., $K^3_4$), and by the analogy that homeomorphs in 3-graphs should appear roughly once spheres (homeomorphs of $S^2$) can be found, just as homeomorphs in graphs appear roughly once cycles (homeomorphs of $S^1$) can be found.
Source paper
A universal exponent for homeomorphs
Peter Keevash, Jason Long, Bhargav Narayanan, Alex Scott · 2020-04-06
https://arxiv.org/abs/2004.02657
PDF source