By OpenAISourceGitHub

Manuscripts from openai/math

Fork
Clone
git clone https://openversions.com/git/index/openai-math.git
Cite
Version link
https://openversions.com/index/openai-math/commit/dfd8c46870b48511e697cbb43a3c8c91d375fa2f
Version identifier (commit hash)
dfd8c46870b48511e697cbb43a3c8c91d375fa2f

Fork to create your own copy with credit to the source. Clone to get a Git copy on your computer.

Latest version

Checks complete

Sync OpenAI Math manuscripts from 4eaee87548250b2682affda7d05d947c05554bae

Committed

All papers

Claims

Findings

The Solomon–Yau least-volume theorem

For every integer m≥2m \ge 2, if MmM^m is a closed connected smooth manifold and F:M→Sm+1F : M \to S^{m+1} is a smooth minimal immersion into the unit round sphere whose image F(M)F(M) is not totally geodesic, then Vol(M,F∗ground)≥am\mathrm{Vol}(M, F^* g_{\mathrm{round}}) \ge a_m, where am=min⁡1≤k<mVolm(Ck,m−k)a_m = \min_{1 \le k < m} \mathrm{Vol}_m(\mathcal{C}_{k, m-k}) is the least volume of an mm-dimensional minimal Clifford product.

For a smooth minimal immersion F:Mm→Sm+1F : M^m \to S^{m+1} of a closed manifold with m≥2m \ge 2, if some image point has qq preimages, then Vol(M,F∗ground)≥qσm\mathrm{Vol}(M, F^* g_{\mathrm{round}}) \ge q \sigma_m; in particular, every immersion violating Theorem 1.1 is an embedding.

Every smooth, closed, connected, embedded minimal hypersurface Σm⊂Sm+1\Sigma^m \subset S^{m+1} (m≥2m \ge 2) that is not totally geodesic satisfies ∫Σ∣A∣2 dV≥mVolm(Σ)\int_\Sigma |A|^2 \, dV \ge m \mathrm{Vol}_m(\Sigma).

Every smooth, closed, connected, embedded minimal hypersurface Σm⊂Sm+1\Sigma^m \subset S^{m+1} (m≥2m \ge 2) that is not totally geodesic has ind(Σ)≥m+3\mathrm{ind}(\Sigma) \ge m + 3; equality ind(Σ)=m+3\mathrm{ind}(\Sigma) = m + 3 holds only if Σ\Sigma is congruent to a minimal Clifford product Sk(k/m)×Sm−k((m−k)/m)S^k(\sqrt{k/m}) \times S^{m-k}(\sqrt{(m-k)/m}) for some 1≤k<m1 \le k < m, and in particular any such hypersurface of volume less than ama_m has ind(Σ)>m+3\mathrm{ind}(\Sigma) > m + 3.

For m≥2m \ge 2, any detecting map Φ:X→Zm0(Sm+1;Z2)\Phi : X \to \mathcal{Z}_m^0(S^{m+1}; \mathbb{Z}_2) on a finite cubical complex (or on RPm+3\mathbb{RP}^{m+3}) with no concentration of mass and D0=sup⁡x∈XM(Φ(x))<∞D_0 = \sup_{x \in X} \mathbf{M}(\Phi(x)) < \infty yields a stationary integral mm-varifold VV in Sm+1S^{m+1} with γm≤∥V∥(Sm+1)≤D0\gamma_m \le \|V\|(S^{m+1}) \le D_0, where γm>σm\gamma_m > \sigma_m depends only on mm; if 2≤m≤62 \le m \le 6, VV can be chosen as a finite sum of smooth closed embedded minimal hypersurfaces with positive integer multiplicities.

For any smooth, closed, connected, embedded, non-totally-geodesic minimal hypersurface Σm⊂Sm+1\Sigma^m \subset S^{m+1} (m≥2m \ge 2), there is a flat-continuous detecting map Φ:RPm+3→Zm0(Sm+1;Z2)\Phi : \mathbb{RP}^{m+3} \to \mathcal{Z}_m^0(S^{m+1}; \mathbb{Z}_2) with no concentration of mass, unique maximum mass ∣Σ∣|\Sigma| at the central parameter, and—whenever ind(Σ)>m+3\mathrm{ind}(\Sigma) > m+3—a flat homotopy to a detecting map Φ~\widetilde{\Phi} with no concentration of mass and sup⁡ξ∈RPm+3M(Φ~(ξ))<∣Σ∣\sup_{\xi \in \mathbb{RP}^{m+3}} \mathbf{M}(\widetilde{\Phi}(\xi)) < |\Sigma|.

Fixing m≥2m \ge 2 and assuming Theorem 1.1 in dimensions 2,…,m−12, \dots, m-1, any nonzero stationary integral kk-dimensional cone C\mathcal{C} in Rk+1\mathbb{R}^{k+1} with Θ(C,0)<ϑm\Theta(\mathcal{C}, 0) < \vartheta_m (and cyclic modulo two when m=2,3m=2,3) is a multiplicity-one hyperplane if 1≤k≤m1 \le k \le m, and is smooth embedded with multiplicity one away from its vertex (with smooth closed connected embedded minimal link in Sm+1S^{m+1}) if k=m+1k = m+1.

Fixing m≥2m \ge 2 and assuming Theorem 1.1 in dimensions 2,…,m−12, \dots, m-1, if the class Em\mathcal{E}_m of smooth closed connected embedded non-totally-geodesic minimal hypersurfaces Σ⊂Sm+1\Sigma \subset S^{m+1} with Vol(Σ)<am\mathrm{Vol}(\Sigma) < a_m is nonempty, then H=inf⁡Σ∈EmVol(Σ)H = \inf_{\Sigma \in \mathcal{E}_m} \mathrm{Vol}(\Sigma) is attained by some Σ∗∈Em\Sigma_* \in \mathcal{E}_m, with σm<γm≤H<am\sigma_m < \gamma_m \le H < a_m.

Methods

Two-sided curvature maximum μ(x) = sup_{y ≠ x} |⟨N(x), y⟩| / (1 - ⟨x, y⟩) estimated via smooth upper-test functions and a Poisson comparison to prove ∫_Σ |A|^2 dV ≥ m Vol_m(Σ).

Source excerpts

“Rather than differentiate the maximum, we estimate smooth upper tests. The logarithmic form of the estimate will allow comparison with a Poisson potential, forcing the mean of m − P to be nonpositive.” · Show in paper

Conformal balancing of coordinate functions against the positive first Jacobi eigenfunction combined with spectral equality to classify least-index minimal hypersurfaces.

Source excerpts

“We combine the curvature lower bound with the conformal energy upper bound. Under the least-index hypothesis, the two estimates force equality in the spectral comparisons and hence parallel second fundamental form.” · Show in paper

Projective cohomological detection via the two-sheeted filling cover C(Sm+1S^{m+1}) → Z_m^0(Sm+1S^{m+1}; Z_2), spherical isoperimetry gap, and cubical replacement for Almgren–Pitts min–max.

Source excerpts

“Lemma 4.3 (Projective criterion). Set Q = Bm+2 × [−1, 1]. Suppose q ↦ Eq ∈ C(Sm+1) is continuous in volume, with E−q = Ecq (q ∈ ∂Q), E(0,−1) = ∅, E(0,1) = Sm+1. Then q ↦ ∂[Eq] descends to a detecting map on Q/(q ∼ −q on ∂Q) ≅ RPp.” · Show in paper

Lorentz-boosted signed Euclidean cone-distance graph family on RP^{m+3} with restricted Jacobian area-formula estimates and local index deformation.

Source excerpts

“The construction below uses the signed Euclidean distance to the cone over Σ. Acting on its graph by Lorentz transformations produces a family whose area can be estimated on the fixed source Σ, including at cut points and at infinite boosts.” · Show in paper

Dimension-inductive low-density cone regularity via radial line splitting, Allard's interior regularity theorem, and White's mod-two cyclic convergence theorem.

Source excerpts

“The regularity argument has two induction parameters: the hypersurface dimension m in Theorem 1.1, and the dimension of a Euclidean cone. Keeping them separate is essential. The theorem in dimensions less than m will classify cones of dimension at most m, and these classifications will give a smooth link for a cone of dimension m + 1.” · Show in paper