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.

All papers

Claims

Findings

Sharp integral fillings in CAT(0) spaces

Let XX be a proper CAT(0) space and n≥2n \ge 2 an integer. Every compactly supported T∈In(X)T \in \mathbf{I}_n(X) with ∂T=0\partial T = 0 admits a compactly supported S∈In+1(X)S \in \mathbf{I}_{n+1}(X) such that ∂S=T\partial S = T and M(S)≤cnM(T)(n+1)/n=M(T)(n+1)/n(n+1)(n+1)/nωn+11/n\mathbf{M}(S) \le c_n \mathbf{M}(T)^{(n+1)/n} = \frac{\mathbf{M}(T)^{(n+1)/n}}{(n+1)^{(n+1)/n} \omega_{n+1}^{1/n}}, with no restriction on the dimension of XX or on the integer multiplicities of the currents, and the coefficient cnc_n is optimal.

Let MM be a Cartan–Hadamard manifold of dimension d≥3d \ge 3. For every bounded domain Ω⊂M\Omega \subset M with smooth boundary, area(∂Ω)≥dωd1/dvol(Ω)(d−1)/d\mathrm{area}(\partial \Omega) \ge d \omega_d^{1/d} \mathrm{vol}(\Omega)^{(d-1)/d}. The same inequality holds for relatively compact sets of finite perimeter, with ambient perimeter in place of boundary area.

Let YY be a compact CAT(0) space. Every integral two-cycle T∈I2(Y)T \in \mathbf{I}_2(Y) admits S∈I3(Y)S \in \mathbf{I}_3(Y) with ∂S=T\partial S = T and M(S)≤c2M(T)3/2=M(T)3/26π\mathbf{M}(S) \le c_2 \mathbf{M}(T)^{3/2} = \frac{\mathbf{M}(T)^{3/2}}{6\sqrt{\pi}}.

For n>2n > 2, assuming the sharp filling theorem in dimension n−1n - 1, for every ϵ∈(0,S∗)\epsilon \in (0, S_*) there is a finite constant Cϵ≥0C_\epsilon \ge 0 independent of uu such that every real Lipschitz function uu on YY satisfies (S∗−ϵ)∥u∥p2≤4βE(u)+Cϵ∥u∥22(S_* - \epsilon)\|u\|_p^2 \le 4\beta E(u) + C_\epsilon \|u\|_2^2, where β=1/(n−2)\beta = 1/(n-2), p=2n/(n−2)p = 2n/(n-2), and S∗=nsn2/nS_* = n s_n^{2/n}.

For n>2n > 2 and the normalized extremizing cycle AA, 0<Λ<S∗0 < \Lambda < S_* and there exists a nonzero nonnegative v∈Hv \in \mathcal{H} minimizing Q(u)/∥u∥p2Q(u)/\|u\|_p^2 over nonzero u∈Hu \in \mathcal{H}, satisfying 0<W:=∫vp dμ=(Λ/n)n/2<sn0 < W := \int v^p \, d\mu = (\Lambda/n)^{n/2} < s_n, Q(v)=nWQ(v) = nW, and the weak Euler equation 4β∫Dv⋅Dψ dμ+n∫vψ dμ=n∫vp−1ψ dμ4\beta \int Dv \cdot D\psi \, d\mu + n \int v\psi \, d\mu = n \int v^{p-1}\psi \, d\mu for all ψ∈H\psi \in \mathcal{H}.

For ν\nu-almost every center yy, lim inf⁡a→∞an/2Jy(a)≥sn2nW\liminf_{a \to \infty} a^{n/2} J_y(a) \ge \frac{s_n}{2^n W}, and consequently Ly≤2(W/sn)2/n<2L_y \le 2(W/s_n)^{2/n} < 2 for ν\nu-almost every yy.

For n>2n > 2 and the minimizer vv, W2≤B∗I(y)W^2 \le B_* I(y) for every y∈Yy \in Y, and in particular ∫YLy dν(y)≥2\int_Y L_y \, d\nu(y) \ge 2.

Methods

Compact reduction of the sharp integral filling problem from a proper CAT(0) space to a compact closed convex ball via 1-Lipschitz radial projection.

Source excerpts

“For x ∈ X, let P(x) be the point of the segment [o, x] at distance min{R, d(o, x)} from o. Geodesics in a CAT(0) space are unique. Comparing two such radial points with their counterparts in the Euclidean comparison triangle, and then using nonexpansion of Euclidean radial projection onto a closed ball, gives d(P(x), P(x′)) ≤ d(x, x′).” · Show in paper

Metric differentiation of integral current charts in CAT(0) spaces using the four-point inequality to establish Euclidean chart norms and exact chart mass formulas.

Source excerpts

“For almost every density point z ∈ Ei, the seminorm in (2.11) is a Euclidean norm. Thus there is a measurable positive definite matrix Gi(z) with Nz(w)2 = wTGi(z)w, Ji(z) = √det Gi(z).” · Show in paper

“For the representation (2.3), let Gi, Ji be as in (2.12). Then ∥C∥ = ∑i (ϕi)#(|θi|Ji dz).” · Show in paper

Radial first variation of an extremizing cycle along geodesics using the whole-current product and quadratic comparison forms.

Source excerpts

“For 0 ≤ t ≤ h, let λ(t, x) = 1 − tg(x), F(t, x) = [y, x]λ(t,x), Ft(x) = F(t, x), where [y, x]a denotes the point at fraction a along the geodesic from y to x.” · Show in paper

Intrinsic coarea inequalities via a countable family of mass tests and radial Euclidean rearrangement using a generalized inverse that accommodates atoms and singular continuous parts.

Source excerpts

“There is a countable collection Tk(Z) of finite lists ((b1, π1), . . . ,(bN , πN )), where the bj are Lipschitz, ∑j |bj | ≤ 1 pointwise, and each πj is a k-tuple of 1-Lipschitz functions, such that every k-current C of finite mass on Z satisfies M(C) = supT ∈Tk(Z) ∑(b,π)∈T C(b, π).” · Show in paper

“Define the generalized inverse for ρ ≥ 0 by h(ρ) = sup{0 ≤ t < U : ρ < R(t)}, sup ∅ := 0.” · Show in paper

Closure of the chart gradient in L^2(mu) and proof of compact inclusion H -> L^2(mu) by projecting whole weighted currents onto R^n and isolating charts in total variation.

Source excerpts

“The proof obtains compactness for projections of whole weighted currents, then isolates charts in total variation.” · Show in paper

Conformally weighted chord distribution comparison via the integral transform J_y(a) = int_Y (1 + a s^2)^{-n} d nu, admissible composite tests in the minimizer equation, and Euclidean tangent density.

Source excerpts

“We study the chord distribution through the transform Jy(a) = ∫Y (1 + as2)−n dν (a ≥ 0).” · Show in paper