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

Critical honeycomb chords with prescribed boundary endpoints

For convex pure lattice hexagons D_R converging after scaling by R to a regular hexagon, and boundary ports a_R, b_R whose scaled limits are distinct points in the relative interiors of sides, the critical chord law satisfies \mathbb{P}_{D_R;a_R,b_R}\{R4/3−εR^{4/3-\varepsilon} \leq L(\gamma) \leq R4/3+εR^{4/3+\varepsilon}\} \to 1 for every \varepsilon > 0. The same conclusion holds in a full half-plane for prescribed boundary ports at distance R, in a strip of height comparable to R for one prescribed port on each opposite boundary cut at mutual distance O(R), and in a wedge of angle \pi/3 or 2\pi/3 formed by two nonparallel pure cuts with one prescribed port on each ray having clearance comparable to R from the other cut and distance O(R) from the apex, with constants uniform on fixed compact sets of these nondegenerate scaled geometries and normalizing mass at least cR^{-5/4}.

For two macroscopically separated side-interior ports in the hexagons of Theorem 1.1, in a parallel strip of thickness comparable to R with one port on each opposite cut and displacement O(R), and in compact nondegenerate wedge geometries with one port on each ray, the critical mass of connecting paths inside the domain is at least cR^{-5/4}, with constants uniform on fixed compact sets of these scaled configurations.

In the hexagons of Theorem 1.1, fixing a source port tending after scaling to a side-interior point and summing terminal ports over a macroscopic interval converging to a compact side-interior interval at positive distance from the source limit gives length R4/3+o(1)R^{4/3+o(1)} in probability under critical weights, with normalizer comparable to R−1/4R^{-1/4}; the same holds when the source is also summed over a separated macroscopic side-interior interval.

Under the critical half-plane law w(\gamma)/K_l for two boundary ports at distance l, \mathbb{P}\{l4/3−εl^{4/3-\varepsilon} \leq L(\gamma) \leq l4/3+εl^{4/3+\varepsilon}\} \to 1 for every \varepsilon > 0, and the same conclusion holds in the hexagons of Theorem 1.1 when both scaled endpoints converge to distinct points in one side interior.

Let B_N be the critical mass of strict bridges from one fixed port to all ports on the opposite pure cut at integral height N. With \Delta f_N = fN+1f_{N+1} - f_N, for every fixed integer \ell \geq 0 and every \eta > 0 there is Cℓ,ηC_{\ell,\eta} < \infty such that |\Delta^\ell B_N| \leq Cℓ,ηC_{\ell,\eta} N−1/4−ℓ+ηN^{-1/4-\ell+\eta} for all N \geq 1.

For the one-piece irreducible bridge height H, for every \eta > 0 there is C_\eta < \infty such that p(H = n) \leq C_\eta n−7/4+ηn^{-7/4+\eta} for all n \geq 1.

For k \geq 1 independent irreducible bridge pieces with displacement sum S_k, there is C < \infty such that \sup_w \Pr(S_k = w) \leq C k−8/3k^{-8/3} and \sum_{m \in \mathbb{Z}} \sup_{H(w)=m} \Pr(S_k = w) \leq C k−4/3k^{-4/3}.

Fixing a bound on the ratio of transverse displacement to height, for compatible prescribed ports on two parallel pure cuts at integral height h \to \infty, normalizing critical bridge weights by their total mass gives \Pr\{h4/3−ϵh^{4/3-\epsilon} \leq L \leq h4/3+ϵh^{4/3+\epsilon}\} \to 1 uniformly for every \epsilon > 0; the same holds in convex pure-sided domains in a fixed compact family of nondegenerate scaled geometries with ports on opposite parallel sides and order-h clearance from every other side.

In a pure wedge of angle \pi/3 or 2\pi/3 with one prescribed port on each boundary ray having order-R clearance to the opposite cut and distance O(R) from the apex, under the critical chord law the length is R4/3+o(1)R^{4/3+o(1)} in probability, and this conclusion persists after restriction to the hexagons of Theorem 1.1 with the same side-interior limiting data.

For two half-plane boundary ports separated by each integer g \to \infty, the probability law \kappa^{L(\gamma)}/K_g satisfies L(\gamma) = g4/3+o(1)g^{4/3+o(1)} in probability and \mathbb{E} L(\gamma) = g4/3+o(1)g^{4/3+o(1)}.

There is one set G \subseteq \mathbb{N} with |G \cap [1, N]|/N \to 1 such that, for every fixed \delta, \varepsilon > 0, all sufficiently large h \in G satisfy h4/3−εh^{4/3-\varepsilon} \leq b_h(L)h^{(L)}(e)/b_h(e) \leq h4/3+εh^{4/3+\varepsilon} for all e \in \mathbb{Z} + h/2 with |e| \leq h1−δh^{1-\delta}.

For E_H = 2\mathbb{N} \cap [H, 2H], Z_H = \sum_{h \in E_H} b_h(0), and M_H = \sum_{h \in E_H} b_h(L)h^{(L)}(0), giving each pair (h, \gamma) with h \in E_H and \gamma \in \mathcal{B}_h(0) probability \kappa^{L(\gamma)}/Z_H yields Z_H = H−1/4+o(1)H^{-1/4+o(1)}, M_H = H13/12+o(1)H^{13/12+o(1)}, and \mathbb{E}_H L = M_H/Z_H = H4/3+o(1)H^{4/3+o(1)} as H \to \infty through all real scales.

Methods

Interior cut-averaging connector construction: approximating smooth interior routes by finitely many segments in pure normal directions with narrow corridors and averaging the splitting indicator over independent junction-cut windows (with one shared terminal window) via Cauchy-Schwarz to obtain a cR^{-5/4} mass lower bound.

Source excerpts

“We adapt the cut-averaging construction in the proof of [16, Proposition 10.2], placing the routes in the interior rather than the exterior. The geometric choices below ensure that the decomposition can be recovered for two different lists of cuts.” · Show in paper

Recoverable translation-and-attachment construction: translating an input chord by lattice vectors v with positive inward normal projections and attaching restricted bridges from the translated endpoints back to the original supports so that the decomposition is uniquely recovered from the crossings.

Source excerpts

“Translate an input chord by a lattice vector v with |v| ≤ Cs and both g1 = n · v/d, g2 = m · v/d between fixed positive multiples of s. Attach a height-gi bridge from each translated endpoint out to its original support.” · Show in paper

Common-level pair-renewal and joining-arch construction: growing two independent sequences of upward irreducibles from two ports on one line, stopping at common positive renewal heights, and joining them above a common level to prove sublinear gap regularity and same-side length concentration.

Source excerpts

“Run two independent sequences of upward irreducibles of law p, translating each new piece to the current endpoint. Stop at their next common positive renewal height.” · Show in paper

Conformal transport of residual spaces to the unit disk and finite-difference analysis of the evaluation kernel and shifted-contour Fourier transforms to bound all fixed forward differences of the strip partition function B_N.

Source excerpts

“The strip | Re ω| < b/2 maps bijectively to the unit disk. The rational function Q is an even finite Blaschke product; its zeros are 0 with multiplicity two and ±v(λ) with multiplicity one.” · Show in paper

Exterior-arc boundary comparison for nonparallel cuts: adding a formal exterior semicircular arc between two target ports P and Q with phase eiσΔWe^{i\sigma\Delta W} and using the local contour identity to compare exceptional masses at P and Q.

Source excerpts

“Add a formal connection from P to Q in the exterior of the target half-plane: a semicircle tangent to the normal rays at its anchors. This arc is not a honeycomb walk. Assign it modulus weight one, no visited-center length, and the phase e iσ∆W of its geometric turn.” · Show in paper

Marked-arm decomposition with deterministic index allocation: splitting a half-plane arch at a highest vertex and allocating disjoint piece blocks to a shared-turn test, a separated-port weak-order test, a length/jump mark, and a two-coordinate match.

Source excerpts

“Take the first ⌊u/16⌋ pieces of each arm for the top test and the last ⌊v/16⌋ for the bottom test. Neither block contains a designated piece. They are disjoint since u + v ≤ ki − 1 for both arms.” · Show in paper

Two-bridge spectator-tripod endpoint variation method: pairing a bridge spectator with an active exit stopped before first contact to express first differences b_h(y'-1) - b_h(y') in terms of two vertex-disjoint bridges.

Source excerpts

“Indeed take a bridge spectator from 1 to y ′ on top and grow an active exit from 0 with phase e iσW , stopping instead just before first contact with the spectator vertices.” · Show in paper