- Boxes
- definitions
- Ellipses
- theorems and lemmas
- Blue border
- the statement of this result is ready to be formalized; all prerequisites are done
- Orange border
- the statement of this result is not ready to be formalized; the blueprint needs more work
- Blue background
- the proof of this result is ready to be formalized; all prerequisites are done
- Green border
- the statement of this result is formalized
- Green background
- the proof of this result is formalized
- Dark green background
- the proof of this result and all its ancestors are formalized
- Dark green border
- this is in Mathlib
The PINN parameterizes perturbations via the 14-dimensional \(\mathfrak {g}_2\) adjoint:
Only 14 functions are learned (not 35).
The trained PINN is evaluated on a grid and FFT identifies dominant modes. Coefficients are rationalized to \(\mathbb {Q}\) within tolerance \(10^{-8}\).
The six hypotheses (H1)–(H6) for spectral bounds:
Volume normalization: \(\mathrm{Vol}(K) = 1\)
Bounded neck volume: \(v_0 \leq \mathrm{Vol}(N) \leq v_1\)
Product metric: \(g|_N = dt^2 + g_Y\)
Block Cheeger bound: \(h(M_i \setminus N) \geq h_0\)
Balanced blocks: \(\mathrm{Vol}(M_i) \in [1/4, 3/4]\)
Neck minimality: \(\mathrm{Area}(\Gamma ) \geq \mathrm{Area}(Y)\)
A TCS manifold \(K = M_1 \cup _N M_2\) consists of:
Two asymptotically cylindrical manifolds \(M_1, M_2\) (building blocks)
A cylindrical neck \(N \cong Y \times [0, L]\) with cross-section \(Y\)
Neck length parameter \(L {\gt} 0\)
Volume normalization: \(\mathrm{Vol}(K) = 1\)
The canonical GIFT G2 metric on \(K_7\) is given by:
3-form (\(35\) components, \(7\) non-zero):
where \(c = (65/32)^{1/14}\).
Metric (\(7 \times 7\) diagonal):
Properties:
\(\det (g) = 65/32\) (exact)
\(\| T\| = 0\) (torsion-free)
\(\mathrm{Hol}(g) = G_2\) (by construction)
Over the first 100,000 zeros: \(\max _n |N_{\text{approx}}(T_n) - n| = 0.111 {\lt} 0.5\) (safety factor \(4.52\times \)).
98.0% of zeros are localized to their correct inter-Gram interval. The 2.0% failure rate is predicted (within 8%) by a simplified GUE model.
The adaptive formula explains 93.9% of the variance in zero corrections \(\delta _n\) (\(R^2 = 0.939\), out-of-sample \(R^2 = 0.919\)).
For \(\varphi = c \cdot \varphi _0\) and metric \(g_{ij} = \frac{1}{6}\sum _{k,l} \varphi _{ikl}\varphi _{jkl}\):
Standard \(\varphi _0\) gives \(g = I_7\), so \(\det (g) = 1\)
Scaling \(\varphi \mapsto c \cdot \varphi \) gives \(g \mapsto c^2 \cdot g\)
Therefore \(\det (g) \mapsto c^{14} \cdot \det (g)\)
Setting \(c^{14} = 65/32\) yields \(\det (g) = 65/32\)
For a compact Riemannian manifold \(M\) and a suitable test function \(h\):
Reference: Selberg (1956); Duistermaat-Guillemin (1975), Invent. Math. 29:39–79.