Palette Sparsification-Type Result

• The paper demonstrates that the uniform Turán density equals the supremum of Lagrangians computed from homomorphism-minimal palettes, ensuring tight bounds.
• The palette sparsification principle simplifies hypergraph analysis by reducing the vast space of palettes to a finite kernel of extremal obstructions.
• The method leverages structural tools like the palette-classification theorem, ordered witnesses, and grid-Ramsey lemmas to ensure computational efficiency.

A palette sparsification-type result, in the context of extremal combinatorics and hypergraph Turán-type problems, provides a conceptual and algorithmic framework for reducing the complexity of dense combinatorial constructions involving colors (“palettes”) without loss of extremality. These results establish that, for determining the uniform Turán density of a given hypergraph, it suffices to consider only a minimal, homomorphism-irreducible family of palettes—thus “sparsifying” the potentially vast space of possible constructions. This concept both extends and unifies lower bound constructions associated with classical palette methods, and delivers a practical reduction principle with direct algorithmic and structural implications for extremal set theory.

1. Palettes and the Lower Bound Construction

A palette is a pair $\mathcal{P} = (C, T)$, where $C$ is a finite color set and $T \subseteq C^3$ is a set of feasible triples—interpreted as allowed triples of colors that may appear on the pairs of a 3-uniform hyperedge. Each palette is associated with two essential invariants:

• Density: $d(\mathcal P) = |T|/|C|^3$,
• Lagrangian: $$L(\mathcal{P}) = \max_{\substack{p\colon C\to[0,1],\,\sum_{x\in C}p(x)=1}} \sum_{(x,y,z)\in T} p(x) p(y) p(z)$$.

Given a palette $\mathcal{P}$, one forms a random 3-uniform hypergraph $H_n$ by coloring each unordered pair of $n$ vertices randomly from $C$, and including each triple as an edge if its collection of pair colors forms a permitted triple in $T$. For large $n$, $H_n$ is $(d(\mathcal{P})-\varepsilon,\varepsilon)$-uniformly dense with high probability for any $\varepsilon>0$, and if a fixed 3-graph $H$ is not $\mathcal{P}$-colorable, then $H$ cannot appear as a subhypergraph of $H_n$. Hence,

$\pi_{\rm uniform}(H) \geq d(\mathcal{P}),$

and even

$\pi_{\rm uniform}(H) \geq L(\mathcal{P}),$

where $\pi_{\rm uniform}(H)$ denotes the uniform Turán density of $H$ (Král' et al., 22 May 2025).

The key result, established via the hypergraph-regularity method, is that these palette bounds are always tight. For every 3-uniform hypergraph $H$,

$$\pi_{\rm uniform}(H) = \sup \{ L(\mathcal{P}) : H \text{ is not } \mathcal{P}\text{-colorable} \}.$$

2. Palette-Classification Theorem

The palette-classification theorem determines whether there exists a 3-uniform hypergraph that is colorable by one palette but not by another. A palette homomorphism $$\varphi : \mathcal{P} = (C, T) \rightarrow \mathcal{Q} = (C', T')$$ is a map $\varphi: C \rightarrow C'$ such that

$$(x, y, z) \in T \implies (\varphi(x), \varphi(y), \varphi(z)) \in T'.$$

Symmetrization $(\mathcal{Q})$ broadens the set $T'$ by including all permutations of triples with appropriately labeled “clones” to account for vertex orderings in colorings.

The core result: there exists a 3-graph $H$ colorable by $\mathcal{P}$ but not by $\mathcal{Q}$ if and only if there is no homomorphism from $\mathcal{P}$ to $\mathcal{Q}$ nor to its symmetrization. This extends to a multipalette setting: given families $\{\mathcal{P}_i\}$ and $\{\mathcal{Q}_j\}$, there exists an $H$ colorable by all $\mathcal{P}_i$ but none of the $\mathcal{Q}_j$ if and only if certain mixed-product palettes admit no homomorphism into any $\mathcal{Q}_j$ or their symmetrizations.

3. Palette Sparsification Principle

The palette-classification and homomorphism framework yields a “palette sparsification” principle: to compute or bound $\pi_{\rm uniform}(H)$, it suffices to consider only palettes that are minimal under the homomorphism partial order. Specifically, if a palette $\mathcal{P}'$ is a homomorphic image of $\mathcal{P}$, the latter is redundant for bounding $\pi_{\rm uniform}(H)$. The supremum in

$$\pi_{\rm uniform}(H) = \sup \{ d(\mathcal{P}_\alpha) : H \text{ not } \mathcal{P}_\alpha\text{-colorable} \}$$

can be restricted to the family of homomorphism-minimal palettes: $$\{ \mathcal{P}_\alpha : \nexists\,\beta \neq \alpha \text{ with } \mathcal{P}_\alpha \rightarrow \mathcal{P}_\beta \}.$$ This reduces the potentially infinite search over palettes to a finite “kernel” of extremal obstructions (Král' et al., 22 May 2025).

4. Application: The 4/81-Example

The sparsification method is illustrated by determining $\pi_{\rm uniform}(H_{4/81}) = 4/81$ for a specific 3-graph. This result is achieved by identifying two small palettes: $\mathcal{P}_{\rm LM}$ and $\mathcal{P}_{3T}$, neither of which admits a homomorphism to a third palette $\mathcal{P}_{4/81}$ (density 4/81), while any palette of greater density does receive a homomorphism from one of these two. Hence, only these two palettes are required to certify extremality, rather than considering the full space of possible palette constructions. This approach extends to other solved cases such as the broken tetrahedron and related 3-graphs.

5. Key Lemmas and Combinatorial Tools

Several combinatorial and structural tools are central to the palette sparsification method:

• Homomorphism ⇒ colorability: Ensures that any $\mathcal{P}$-colorable $H$ is also $\mathcal{Q}$-colorable if a homomorphism exists ($\mathcal{P} \to \mathcal{Q}$).
• Existence of ordered witnesses: Compact Ramsey-regularity arguments construct small “witness” hypergraphs for colorability obstructions.
• Sparsification via grid-Ramsey and ES-grid lemmas: These Ramsey-type results provide efficient reductions in the multipalette case by identifying factors through grid colorings.
• Sunflower-type probabilistic constructions and Chernoff concentration bounds: Guarantee with high probability that random constructions exhibit the required structural sparsity and colorability/non-colorability properties.

6. Computational Strategy and Implications

To compute or bound $\pi_{\rm uniform}(H)$ in practice, one:

• Identifies a finite set of candidate minimal palettes based on known extremal configurations.
• Verifies that no palette in this set maps homomorphically to another, ensuring minimality.
• Computes their Lagrangians to find the maximal value.
• Uses the palette-classification theorem to guarantee that any other palette that might avoid $H$ must have smaller or equal Lagrangian.

This methodological reduction bypasses the need for deep hypergraph-regularity arguments in most cases, replacing them with finite, checkable palette homomorphism computations. The palette sparsification-type result thereby offers a powerful and conceptually transparent method for determining uniform Turán densities and understanding the structure of extremal hypergraphs, with reduction principles critical for both theoretical investigations and algorithmic applications (Král' et al., 22 May 2025).