OctoInformBeta

OpenAI preprint, September 2026 · §3. A concave cometric and smooth product cores

Lemma 3.1 [Uniform cometric estimates]: For a sufficiently small positive neck parameter δ\delta, the estimates hold in the original and both endpoint frames

Uniform bounds for the neck cometric.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

The neck cometric is a family of quadratic forms on the coordinate momentum plane ℝ2\mathbb R^2, parametrized by the neck parameter δ\delta and height tt. The endpoint frames use inward height τ\tau.

For a coordinate momentum p=(p1,p2)𝖳∈ℝ2p=(p_1,p_2)^{\mathsf T}\in\mathbb R^2 and a real 2×22\times2 matrix HH, the squared Euclidean norm and the quadratic form of HH are defined by (1): |p|2=p12+p22,p𝖳Hp=∑i,j=12piHijpj.|p|^2=p_1^2+p_2^2, \qquad p^{\mathsf T}Hp=\sum_{i,j=1}^2 p_i H_{ij}p_j.(1) The Euclidean matrix operator associated with HH is the linear map p↦Hpp\mapsto Hp; its operator norm is denoted by ∥H∥op\|H\|_{\mathrm{op}}.✦ Matrix inequalities mean inequalities of quadratic forms, and Id\mathrm{Id} denotes the 2×22\times2 identity matrix.

The Cometric and Its Endpoint Frames

The one-sided cutoff a:ℝ→ℝa:\mathbb R\to\mathbb R is the smooth exponential cutoff in (2); the integrated cutoff A:ℝ→ℝA:\mathbb R\to\mathbb R is its integral from zero: a(u)={exp⁡(−12/3−u),u<2/3,0,u≥2/3,A(u)=∫0ua(y)dy.a(u)= \begin{cases} \exp\!\left(-\dfrac{1}{2/3-u}\right),&u<2/3,\\ 0,&u\ge 2/3, \end{cases} \qquad A(u)=\int_0^u a(y)\,dy.(2)

Let θ:ℝ→ℝ\theta:\mathbb R\to\mathbb R be the standard smooth transition from zero to one. The shear transition χ:ℝ→ℝ\chi:\mathbb R\to\mathbb R and the cometric shear σ:ℝ2→ℝ\sigma:\mathbb R^2\to\mathbb R are the functions defined by (3): χ(t)=θ(6(t−512)),σ(δ,t)=δ2χ(t).\chi(t)=\theta\!\left(6\left(t-\frac{5}{12}\right)\right), \qquad \sigma(\delta,t)=\delta^2\chi(t).(3)

The diagonal coefficients d1,d2:ℝ2→ℝd_1,d_2:\mathbb R^2\to\mathbb R are the functions in (4): d1(δ,t)=1+δA(t),d2(δ,t)=1+δA(1−t).d_1(\delta,t)=1+\delta A(t), \qquad d_2(\delta,t)=1+\delta A(1-t).(4) At parameter δ\delta and height tt, the cometric matrix P(δ,t)P(\delta,t) is the real 2×22\times2 matrix in (5), and its metric matrix Q(δ,t)Q(\delta,t) is its inverse: P(δ,t)=(d1(δ,t)+d2(δ,t)σ(δ,t)2d2(δ,t)σ(δ,t)d2(δ,t)σ(δ,t)d2(δ,t)),Q(δ,t)=P(δ,t)−1.\begin{aligned} P(\delta,t)&= \begin{pmatrix} d_1(\delta,t)+d_2(\delta,t)\sigma(\delta,t)^2 &d_2(\delta,t)\sigma(\delta,t)\\ d_2(\delta,t)\sigma(\delta,t)&d_2(\delta,t) \end{pmatrix},\\ Q(\delta,t)&=P(\delta,t)^{-1}. \end{aligned}(5)

The endpoint frame matrix C2(δ)C_2(\delta) and the right endpoint cometric Pend(δ,τ)P_{\mathrm{end}}(\delta,\tau) are the real 2×22\times2 matrices defined in (6): C2(δ)=(011−δ2),Pend(δ,τ)=C2(δ)P(δ,1−τ)C2(δ)𝖳.\begin{aligned} C_2(\delta)&=\begin{pmatrix}0&1\\1&-\delta^2\end{pmatrix},\\ P_{\mathrm{end}}(\delta,\tau) &=C_2(\delta)P(\delta,1-\tau)C_2(\delta)^{\mathsf T}. \end{aligned}(6) The endpoint coordinate cometrics R0,R1:ℝ2→ℝ2×2R_0,R_1:\mathbb R^2\to\mathbb R^{2\times2} are the matrix-valued functions in (7), with 00 denoting the left frame and 11 the right frame: R0(δ,τ)=P(δ,τ),R1(δ,τ)=Pend(δ,τ).R_0(\delta,\tau)=P(\delta,\tau), \qquad R_1(\delta,\tau)=P_{\mathrm{end}}(\delta,\tau).(7)✦ Primes on PP denote derivatives with respect to tt; primes on RrR_r denote derivatives with respect to τ\tau, for r∈{0,1}r\in\{0,1\}.

The Claim

Theorem 1 (Uniform cometric estimates). There are constants 0<δo≤1/20<\delta_{\mathrm{o}}\le 1/2 and Co>0C_{\mathrm{o}}>0 such that, whenever 0<δ≤δo0<\delta\le\delta_{\mathrm{o}} and t∈[0,1]t\in[0,1], the original cometric satisfies (8) for every p∈ℝ2p\in\mathbb R^2: |p|23≤p𝖳P(δ,t)p≤6|p|2,|p|26≤p𝖳Q(δ,t)p≤3|p|2,|P′(δ,t)p|2≤Co2δ2|p|2,p𝖳(−P″(δ,t))p≥0,|P′(δ,t)p|2≤Coδp𝖳(−P″(δ,t))p.\begin{aligned} \frac{|p|^2}{3} &\le p^{\mathsf T}P(\delta,t)p\le 6|p|^2,\\ \frac{|p|^2}{6} &\le p^{\mathsf T}Q(\delta,t)p\le 3|p|^2,\\ |P'(\delta,t)p|^2 &\le C_{\mathrm{o}}^2\delta^2|p|^2,\\ p^{\mathsf T}\bigl(-P''(\delta,t)\bigr)p&\ge0,\\ |P'(\delta,t)p|^2 &\le C_{\mathrm{o}}\delta\, p^{\mathsf T}\bigl(-P''(\delta,t)\bigr)p. \end{aligned}(8)✦

For the endpoint coordinate cometrics, there are constants 0<δe≤1/20<\delta_{\mathrm{e}}\le1/2 and Ce>0C_{\mathrm{e}}>0, independent of the endpoint r∈{0,1}r\in\{0,1\}, such that, whenever 0<δ≤δe0<\delta\le\delta_{\mathrm{e}}, τ∈[0,1]\tau\in[0,1], and p∈ℝ2p\in\mathbb R^2, the estimates (9) hold: |p|29≤p𝖳Rr(δ,τ)p≤18|p|2,|p|218≤p𝖳Rr(δ,τ)−1p≤9|p|2,|Rr′(δ,τ)p|2≤Ce2δ2|p|2,p𝖳(−Rr″(δ,τ))p≥0,|Rr′(δ,τ)p|2≤Ceδp𝖳(−Rr″(δ,τ))p.\begin{aligned} \frac{|p|^2}{9} &\le p^{\mathsf T}R_r(\delta,\tau)p\le18|p|^2,\\ \frac{|p|^2}{18} &\le p^{\mathsf T}R_r(\delta,\tau)^{-1}p\le9|p|^2,\\ |R_r'(\delta,\tau)p|^2 &\le C_{\mathrm{e}}^2\delta^2|p|^2,\\ p^{\mathsf T}\bigl(-R_r''(\delta,\tau)\bigr)p&\ge0,\\ |R_r'(\delta,\tau)p|^2 &\le C_{\mathrm{e}}\delta\, p^{\mathsf T}\bigl(-R_r''(\delta,\tau)\bigr)p. \end{aligned}(9)✦

There are also constants 0<δop≤1/20<\delta_{\mathrm{op}}\le1/2 and Cop>0C_{\mathrm{op}}>0 such that, whenever 0<δ≤δop0<\delta\le\delta_{\mathrm{op}} and τ∈[0,1]\tau\in[0,1], the operator estimate (10) holds at either endpoint: ∥Rr′(δ,τ)∥op≤Copδ,r∈{0,1}.\bigl\|R_r'(\delta,\tau)\bigr\|_{\mathrm{op}} \le C_{\mathrm{op}}\delta, \qquad r\in\{0,1\}.(10)✦

In the paper, the result reads: there are δ0>0\delta_0>0 and C<∞C<\infty, depending only on the fixed χ\chi and independent of the parameters B1,B2B_1,B_2, such that for 0<δ≤δ00<\delta\le\delta_0, in the original coordinates and either endpoint coordinate system, Id/9≤P≤18Id\mathrm{Id}/9\le P\le18\,\mathrm{Id}, Id/18≤Q=P−1≤9Id\mathrm{Id}/18\le Q=P^{-1}\le9\,\mathrm{Id}, ∥P′∥op≤Cδ\|P'\|_{\mathrm{op}}\le C\delta, −P″≥0-P''\ge0, and |P′p|2≤Cδp𝖳(−P″)p|P'p|^2\le C\delta\,p^{\mathsf T}(-P'')p for every p∈ℝ2p\in\mathbb R^2, where ∥⋅∥op\|\cdot\|_{\mathrm{op}} is the Euclidean operator norm.

What Is Checked

The assertions concern the explicit real 2×22\times2 cometric PP built from AA and χ\chi, its inverse QQ, and the endpoint coordinate cometrics R0,R1R_0,R_1, for sufficiently small positive δ\delta and heights in [0,1][0,1].