\documentclass[11pt,a4paper]{article} \usepackage{amsmath,amssymb,amsthm} \usepackage{geometry} \usepackage{hyperref} \usepackage{enumitem} \usepackage{mathrsfs} \geometry{margin=2.5cm} \newtheorem{theorem}{Theorem} \newtheorem{lemma}[theorem]{Lemma} \newtheorem{proposition}[theorem]{Proposition} \newtheorem{corollary}[theorem]{Corollary} \theoremstyle{definition} \newtheorem{definition}{Definition} \newtheorem{conjecture}{Conjecture} \newtheorem{remark}{Remark} \title{\textbf{Research Roadmap: Toward a Mathematical Physics Paper}\\[6pt] \large Low-Dimensional Invariant Structure of the Balanced Metric\\ on the Dwork Quintic Family} \author{D. Yang Eng \\ \small Internal research notes} \date{\today} \begin{document} \maketitle \tableofcontents \bigskip % ------------------------------------------------------------------ \section{The Central Mathematical Question} % ------------------------------------------------------------------ The empirical paper established the following observation: \begin{quote} \textit{For the Fermat quintic $X_0 \subset \mathbb{CP}^4$ with the $k=10$ balanced metric, the determinant ratio $R_\psi(z) = \log(\det g_k / \det g_{\mathrm{FS}})$ is a function effectively of only two symmetric invariants $(p_2, e_3)$, with all remaining independent generators contributing $\Delta R^2 < 0.001$.} \end{quote} \noindent The mathematical physics paper asks: \textbf{why is this true, and can it be proved?} This is a genuine open question at the intersection of complex geometry, representation theory, and spectral theory. The empirical evidence is strong. The proof requires understanding how the Monge-Amp\`ere constraint selects specific directions in the ring of $G$-invariant functions on $X_0$. % ------------------------------------------------------------------ \section{Mathematical Setup} % ------------------------------------------------------------------ \subsection{The Fermat and Dwork Quintics} \begin{definition} The \textbf{Dwork family} is the one-parameter family of quintic hypersurfaces \[ X_\psi = \Bigl\{z \in \mathbb{CP}^4 \;\Big|\; \sum_{i=0}^4 z_i^5 - 5\psi\,z_0 z_1 z_2 z_3 z_4 = 0 \Bigr\}, \qquad \psi \in \mathbb{C} \setminus \{1, \omega, \ldots\}. \] The \textbf{Fermat quintic} $X_0$ is the case $\psi = 0$. \end{definition} \subsection{The Symmetry Group} \begin{definition} Let $G = (U(1)^5/U(1)) \rtimes S_5 \cong U(1)^4 \times S_5$ be the \textbf{residual symmetry group} of $X_0$, acting on $\mathbb{CP}^4$ by \[ (e^{i\theta_0}, \ldots, e^{i\theta_4}) : [z_0 : \cdots : z_4] \mapsto [e^{i\theta_0} z_0 : \cdots : e^{i\theta_4} z_4] \qquad (\textstyle\sum_j \theta_j = 0), \] and $S_5$ permutes the coordinates. The Fermat quintic $X_0$ is $G$-invariant. \end{definition} \subsection{The Normalisation Locus and Invariant Generators} Working on the normalisation locus $\Sigma = \{z : \sum_i |z_i|^2 = 1\} \cap X_0$, set $u_i = |z_i|^2$. The $G$-invariant real-valued functions of $(u_0,\ldots,u_4)$ form a polynomial ring generated by the power sums \[ p_k = \sum_{i=0}^4 u_i^k, \qquad k = 2, 3, 4, 5, \] (note: $p_1 = \sum u_i = 1$ is fixed by normalisation). Equivalently, via Newton's identities, one may use the elementary symmetric polynomials $e_2, e_3, e_4, e_5$ as generators, with \[ e_2 = \tfrac{1-p_2}{2}, \quad e_3 = \tfrac{1 - 3p_2 + 2p_3}{6}, \quad e_4 = \tfrac{1 - 6p_2 + 8p_3 + 3p_2^2 - 6p_4}{24}, \quad e_5 = u_0 u_1 u_2 u_3 u_4. \] The \textbf{invariant ring} is thus \[ \mathcal{R} = \mathbb{R}[p_2, p_3, p_4, p_5] = \mathbb{R}[e_2, e_3, e_4, e_5], \] with the two generating sets related algebraically. \subsection{The Balanced Metric and Determinant Ratio} Donaldson's balanced metric $g_k$ at polynomial degree $k$ on $(X_0, \mathcal{O}(k)|_{X_0})$ is the unique Hermitian metric satisfying the balanced condition \[ \frac{1}{N_k} \sum_{\alpha} |s^\alpha(z)|^2_{h_k^{\otimes k}} = \frac{1}{\mathrm{Vol}(X_0)} \] for an $L^2$-orthonormal basis $\{s^\alpha\}$ of $H^0(X_0, \mathcal{O}(k))$, with $N_k = \dim H^0(X_0, \mathcal{O}(k))$. The metric is obtained as the fixed point of Donaldson's $T$-operator. The \textbf{determinant ratio} is \[ R_k(z) \;=\; \log\frac{\det g_k(z)}{\det g_{\mathrm{FS}}(z)}, \] where $\det$ denotes the determinant of the $3 \times 3$ local metric tensor. % ------------------------------------------------------------------ \section{What Can Be Proved Now: The Easy Part} % ------------------------------------------------------------------ The following are rigorous or near-rigorous statements that require only standard arguments. \begin{proposition}[G-invariance of the det-ratio] \label{prop:ginv} The function $R_k : X_0 \to \mathbb{R}$ is $G$-invariant. \end{proposition} \begin{proof}[Proof sketch] The Fermat quintic $X_0$ is $G$-invariant. The Fubini-Study metric $g_{\mathrm{FS}}$ is $G$-invariant. The balanced metric $g_k$ is the unique solution to the balanced equation on $(X_0, \mathcal{O}(k))$; since the equation and the variety are both $G$-invariant, the solution must be too. Therefore $\det g_k / \det g_{\mathrm{FS}}$ is $G$-invariant. \end{proof} \begin{corollary}[Four-generator reduction] \label{cor:four} There exists a smooth function $F_k : \mathbb{R}^4 \to \mathbb{R}$ such that \[ R_k(z) = F_k(p_2, e_3, e_4, e_5) \] for all $z \in X_0$. \end{corollary} \begin{proof} Follows from Proposition~\ref{prop:ginv} and the fact that $\{p_2, e_3, e_4, e_5\}$ (equivalently $\{p_2, p_3, p_4, p_5\}$) generate the $G$-invariant polynomial ring on $\Sigma$. \end{proof} \begin{remark} Corollary~\ref{cor:four} alone reduces the problem from $875^2$ parameters (the full H-matrix space) to a function of 4 real variables. The real mathematical challenge is the further reduction from 4 generators to 2. \end{remark} \begin{lemma}[H-matrix block structure at $\psi=0$] \label{lem:block} At the Fermat point $\psi = 0$, the balanced H-matrix is diagonal in the monomial basis and the diagonal entries $h_\alpha$ depend only on the partition type $\lambda(\alpha) = (\alpha_0 \geq \alpha_1 \geq \cdots \geq \alpha_4)$ of the multi-index $\alpha$. \end{lemma} \begin{proof} The $U(1)^4 \subset G$ action makes the H-matrix diagonal. The $S_5 \subset G$ action makes the diagonal entries depend only on the unordered multi-set of exponents, i.e., the partition type. \end{proof} \begin{remark} For $k = 10$, the number of distinct partition types of 10 into at most 5 parts is exactly 30. So the balanced H-matrix at $\psi=0$ is determined by \emph{30 real parameters}, not 875. The empirical observation that $R_k$ depends on 2 generators rather than 4 is a further reduction on top of this. \end{remark} % ------------------------------------------------------------------ \section{The Main Conjecture: Precise Statements} % ------------------------------------------------------------------ We state three versions, from weakest to strongest. \begin{conjecture}[Weak: approximate two-dimensionality] \label{conj:weak} For the Fermat quintic $X_0$ with the degree-$k$ balanced metric ($k = 6, 8, 10$), the function $F_k : \mathbb{R}^4 \to \mathbb{R}$ from Corollary~\ref{cor:four} satisfies \[ \left\|\frac{\partial F_k}{\partial e_4}\right\|_{L^2(X_0)},\; \left\|\frac{\partial F_k}{\partial e_5}\right\|_{L^2(X_0)} \;\ll\; \left\|\frac{\partial F_k}{\partial p_2}\right\|_{L^2(X_0)}. \] Quantitatively: the variance explained by $(p_2, e_3)$ exceeds 99\% of the total, i.e., $\mathrm{Var}[R_k - \Pi_{(p_2,e_3)} R_k] / \mathrm{Var}[R_k] < 0.001$. \end{conjecture} \begin{conjecture}[Medium: asymptotic two-dimensionality] \label{conj:medium} As $k \to \infty$, the normalised correction $k(R_k - \bar R_k)$, where $\bar R_k$ is the mean value of $R_k$, converges to a function $f_\infty(p_2, e_3)$ depending only on $(p_2, e_3)$: \[ \sup_{z \in X_0} \left| k(R_k(z) - \bar R_k) - f_\infty(p_2(z), e_3(z)) \right| = O(k^{-1}). \] \end{conjecture} \begin{conjecture}[Strong: spectral selection] \label{conj:strong} Let $\mathcal{L}$ be the linearisation of the Monge-Amp\`ere operator around $g_{\mathrm{FS}}|_{X_0}$. The leading eigenfunctions of $\mathcal{L}$ with non-trivial $G$-invariant projection lie in the span of $\{p_2, e_3\}$ (viewed as multiplication operators on $L^2(X_0, g_{\mathrm{FS}})$). \end{conjecture} \medskip \noindent\textbf{Priority.} Conjecture~\ref{conj:weak} is most directly supported by the data and most accessible. Conjecture~\ref{conj:medium} would be a clean mathematical theorem. Conjecture~\ref{conj:strong} would explain the \emph{why}. % ------------------------------------------------------------------ \section{Proof Strategy} % ------------------------------------------------------------------ \subsection{Step 1: Compute the Scalar Curvature of the FS Metric on $X_0$} \textbf{Goal.} Show that the scalar curvature $s(z)$ of the induced FS metric on the Fermat quintic is, or is well-approximated by, a function of $(p_2, e_3)$ alone. \textbf{Why this matters.} The Tian-Yau-Zelditch (TYZ) expansion states that the Bergman kernel satisfies \[ k^{-3} B_k(z,z) \;\sim\; 1 + \frac{s(z)}{2k} + \frac{a_2(z)}{k^2} + \cdots \] as $k \to \infty$. The leading correction to the balanced metric (and hence to $R_k$) is controlled by $s(z)$. If $s(z) = s(p_2, e_3)$, this would imply Conjecture~\ref{conj:medium}. \textbf{How to compute $s$.} For a smooth hypersurface $X \hookrightarrow \mathbb{CP}^n$ of degree $d$ with the induced FS metric, the scalar curvature is \[ s_X = n(n+2) - \frac{(d-1) \cdot |II|^2_{g_{\mathrm{FS}}}}{1}, \] (schematically) where $|II|^2$ is the squared norm of the second fundamental form. For the Fermat quintic $F = \sum_{i=0}^4 z_i^5$, the relevant quantities are: \begin{align} \partial_i F &= 5 z_i^4, \\ \partial_i \partial_j F &= 20 z_i^3 \delta_{ij}, \\ |\nabla F|^2_{g_{\mathrm{FS}}} &= 25 \sum_i |z_i|^8 / (\sum_i |z_i|^2)^4 = 25\, p_4 \quad (\text{on } \Sigma). \end{align} The second fundamental form has components $II_{ij} = \partial_i \partial_j F / |\nabla F|$, which after careful computation gives \[ |II|^2 \;\propto\; \frac{\sum_i |\partial_i \partial_i F|^2}{|\nabla F|^2} = \frac{400 \sum_i |z_i|^6}{25 \sum_i |z_i|^8} = 16 \cdot \frac{p_3}{p_4}. \] (The off-diagonal terms vanish for the Fermat quintic by the diagonal structure $\partial_i \partial_j F = 20 z_i^3 \delta_{ij}$.) \textbf{Key observation.} The scalar curvature involves $p_3 / p_4$. Using $p_3 = (6e_3 - 1 + 3p_2)/2$ (Newton's identity), we can express $p_3$ in terms of $(p_2, e_3)$. However, $p_4$ remains. The question becomes: \begin{quote} \textit{Is $p_4$ effectively a function of $(p_2, e_3)$ in the regime sampled by the Fermat quintic, or does it vary independently?} \end{quote} This must be checked numerically on the actual point cloud. If $p_4$ varies substantially independently of $(p_2, e_3)$ but $R_k$ does not depend on it, then the mechanism is more subtle than just ``scalar curvature controls everything.'' \subsection{Step 2: Spectral Analysis of the Linearised Operator} \textbf{Goal.} Identify which $G$-invariant functions on $X_0$ are in the leading eigenspaces of the Lichnerowicz-type operator that controls the balanced metric iteration. Donaldson's $T$-operator linearised around the balanced metric satisfies \[ \delta(T(H))_{\alpha\beta} = \int_{X_0} K(z,w) \delta H_{\alpha\beta}\, d\mu, \] for some integral kernel $K$. The convergence of the balanced iteration is governed by the spectral gap of this operator, specifically by how quickly the ``non-FS directions'' in H-matrix space are suppressed. The G-invariant sector decomposes into representations of $S_5$. The trivial irrep (symmetric tensors) is spanned by functions of $(p_2, p_3, p_4, p_5)$. \textbf{The conjecture is that the eigenfunctions with large eigenvalues (slow convergence) correspond to $(p_2, e_3)$, while those corresponding to $(e_4, e_5)$ have smaller eigenvalues.} \textbf{Concrete computation.} On $X_0$ with the FS metric, compute the spectrum of the Laplace-Beltrami operator $\Delta_{g_{\mathrm{FS}}}$ restricted to $G$-invariant functions. The basis functions $p_2, e_3, e_4, e_5$ are all in the $G$-invariant sector. If the eigenvalues satisfy $\lambda_2 \sim \lambda_{e_3} \gg \lambda_{e_4} \sim \lambda_{e_5}$, this would suggest $(p_2, e_3)$ are the ``slow modes'' that persist in the balanced metric correction. \subsection{Step 3: Bergman Kernel Asymptotics on the G-Invariant Sector} \textbf{Goal.} Compute the leading TYZ coefficient restricted to the $G$-invariant sector and show it lies in $\mathrm{span}(p_2, e_3)$. The Bergman kernel on $X_0$ decomposes into $G$-irrep sectors. In the trivial (G-invariant) sector: \[ B_k^G(z,z) = \sum_{\alpha \in G\text{-orbits}} \frac{|\mathcal{O}_\alpha(z)|^2}{h_\alpha} \] where $\mathcal{O}_\alpha$ denotes the symmetrised monomial orbit sum. These orbit sums are precisely the $G$-invariant sections of $\mathcal{O}(k)|_{X_0}$, and they are naturally expressible as polynomial functions of $(p_2, p_3, p_4, p_5)$. The key result to prove: the ratio $B_k^G(z,z) / B_k^{G,\mathrm{FS}}(z,z)$, where $B_k^{G,\mathrm{FS}}$ is the $G$-invariant Bergman kernel for the FS metric, is to leading order in $k$ a function of $(p_2, e_3)$ only. \subsection{Step 4: Connect to the Dwork Deformation} For the deformed quintic $X_\psi$, the symmetry group reduces to $\Gamma_\psi \subset G$. The coefficient trajectories $c_i(\psi)$ encode how the dominant modes ($p_2, e_3$) acquire $\psi$-dependent amplitudes. A weaker but still interesting result: show that the deformation vector field $\partial_\psi g_k$ lies in the $(p_2, e_3)$ sector to leading order, using Kodaira-Spencer deformation theory. % ------------------------------------------------------------------ \section{Key Technical Lemmas You Need} % ------------------------------------------------------------------ Ordered by difficulty (ascending): \begin{enumerate}[leftmargin=*] \item \textbf{[Easy, 1 week]} \textit{The $G$-invariant ring on $\Sigma \cap X_0$ is generated by $(p_2, e_3, e_4, e_5)$ as a $\mathbb{R}$-algebra.} This is a standard invariant-theory result. On the normalisation locus $\sum u_i=1$, the ring $\mathbb{R}[u_0,\ldots,u_4]^{S_5}$ is generated by $e_2, e_3, e_4, e_5$ (with $e_1 = 1$ fixed). The U(1)$^4$ action adds no further constraints since it acts on the phases of $z_i$, not on $u_i = |z_i|^2$. \\ \textit{References:} Stanley, ``Invariants of Finite Groups,'' Ch.~3. \item \textbf{[Easy-Medium, 2 weeks]} \textit{Explicit computation of the second fundamental form of $X_0 \hookrightarrow \mathbb{CP}^4$ in coordinates.} Using the formula $II = \nabla^2 F / |\nabla F|$ restricted to the tangent bundle of $X_0$, compute $|II|^2$ explicitly and express it as a function of $(u_0, \ldots, u_4)$. Identify which symmetric polynomials appear. \\ \textit{References:} Griffiths-Harris, ``Principles of Algebraic Geometry,'' Ch.~1.5; Huybrechts, ``Complex Geometry,'' Ch.~3. \item \textbf{[Medium, 4-6 weeks]} \textit{The scalar curvature of $g_{\mathrm{FS}}|_{X_0}$ expressed in terms of $(p_2, p_3, p_4)$.} Use the Gauss equation $s_{X_0} = s_{\mathbb{CP}^4}|_{X_0} - |II|^2$ and the computation from Lemma~2 to get $s$ explicitly. Then use Newton's identities to express $p_3$ in terms of $(p_2, e_3)$ and $p_4$ in terms of $(p_2, e_3, e_4)$. Determine numerically whether the $e_4$ term is negligible on typical points of $X_0$. \\ \textit{References:} Jost, ``Riemannian Geometry and Geometric Analysis,'' Ch.~4. \item \textbf{[Medium, 4-6 weeks]} \textit{First TYZ coefficient for $(X_0, g_{\mathrm{FS}})$.} Verify that the first TYZ coefficient $a_1(z) = s(z)/2$ restricted to $G$-invariant functions on $X_0$ is well-approximated by a function of $(p_2, e_3)$. This is the key analytic lemma connecting the Bergman kernel asymptotics to the observed two-dimensionality. \\ \textit{References:} Zelditch (1998), J. Diff. Geom.; Lu (2000), J. Geom. Anal.; Fine (2012), J. Diff. Geom. \item \textbf{[Hard, 3+ months]} \textit{Spectral gap estimate: the Laplacian on $G$-invariant functions on $(X_0, g_{\mathrm{FS}})$ has a spectral gap separating $(p_2, e_3)$ eigenfunctions from $(e_4, e_5)$ eigenfunctions.} This is the main technical theorem. It requires: (a) computing or estimating the first few eigenvalues of $\Delta_{g_{\mathrm{FS}}}|_{G\text{-inv}}$ on $X_0$, and (b) showing these correspond to the $(p_2, e_3)$ sector. Can potentially be approached numerically first (using the balanced H-matrix data you already have) then rigorously. \\ \textit{References:} Braun-Brelidze-Douglas-Ovrut (2008), JHEP; Ashmore-He-Heyes-Ovrut (2023), JHEP. \item \textbf{[Hard, 3+ months]} \textit{The leading $G$-invariant correction to the balanced metric from the FS metric at large $k$ lies in $\mathrm{span}(p_2, e_3)$.} This is Conjecture~\ref{conj:medium} made rigorous. Approach via the Donaldson iteration as a fixed-point problem and show that the deviation from the FS metric, in the $G$-invariant sector, contracts onto the $(p_2, e_3)$ subspace. \\ \textit{References:} Donaldson (2001), JDG; Fine (2012), JDG; Keller (2009), arXiv. \end{enumerate} % ------------------------------------------------------------------ \section{The Tractable Version of the Paper} % ------------------------------------------------------------------ You do \emph{not} need to prove all six lemmas to write a mathematical physics paper. A publishable paper could contain: \begin{itemize} \item \textbf{Theorem} (easy): $R_k(z)$ is $G$-invariant, hence a function of $(p_2, e_3, e_4, e_5)$ (Lemmas 1). \item \textbf{Computation} (medium): The scalar curvature $s(z)$ of $g_{\mathrm{FS}}|_{X_0}$ expressed explicitly in symmetric polynomials (Lemma 3). Identify whether $e_4, e_5$ appear in $s(z)$. \item \textbf{Conjecture with evidence}: The correction $k(R_k - \bar R_k)$ converges to a function of $(p_2, e_3)$ alone, supported by the polynomial ablation data and k-stability table. \item \textbf{Corollary}: The symbolic formula is the leading-order expression for this limit in the class of rational functions of $(p_2, e_3)$. \end{itemize} \noindent This structure is publishable in \textit{Letters in Mathematical Physics} (with a theorem + conjecture + evidence) or \textit{Journal of Geometry and Physics} (computation + conjecture). % ------------------------------------------------------------------ \section{The Ideal Target Structure} % ------------------------------------------------------------------ If Lemma 4 (TYZ coefficient computation) succeeds and the first TYZ coefficient is indeed a function of $(p_2, e_3)$, then the paper has: \begin{itemize} \item \textbf{Theorem}: First TYZ coefficient on $X_0$ is in $\mathrm{span}(p_2, e_3)$. This implies $R_k \sim f(p_2, e_3) + O(k^{-2})$ for large $k$. \item \textbf{Symbolic formula as corollary}: The SR formula is the rational function of $(p_2, e_3)$ that best approximates this leading correction. \item \textbf{Numerical confirmation}: The ablation tables confirm the $O(k^{-2})$ suppression of $e_4, e_5$ contributions. \end{itemize} This would be a \textit{Communications in Mathematical Physics} paper. % ------------------------------------------------------------------ \section{Immediate Next Steps (Ordered)} % ------------------------------------------------------------------ \begin{enumerate}[leftmargin=*] \item \textbf{This week.} Numerically verify: does $p_4$ vary independently of $(p_2, e_3)$ on your sampled points? Run \texttt{poly3\_r2([p4], R\_psi)} and \texttt{poly3\_r2([p2, e3, p4], R\_psi)} on your Kaggle data. If $p_4$ can be approximated by a function of $(p_2, e_3)$ on $X_0$, then the scalar curvature story is consistent. If not, the mechanism is more subtle. \item \textbf{Week 2-3.} Prove Lemma 1 (G-invariant ring). Write it up carefully with full algebraic proof. This is publishable as-is as a standalone lemma. \item \textbf{Week 3-6.} Compute $|II|^2$ explicitly (Lemma 2). Use the Fermat quintic gradient $\nabla F = 5(z_0^4, \ldots, z_4^4)$ and the Hessian $\nabla^2 F = 20 \, \mathrm{diag}(z_0^3, \ldots, z_4^3)$. Write down $s(z)$ as a function of $(p_2, p_3, p_4)$ explicitly. \item \textbf{Month 2.} Read the TYZ literature (Zelditch 1998, Lu 2000). Compute or look up the second asymptotic coefficient for degree-$d$ hypersurfaces in $\mathbb{CP}^n$. Try to apply it to the quintic case. \item \textbf{Month 3+.} Attempt the spectral gap computation (Lemma 5). Start numerically using the Laplacian eigenvalue code from Braun et al.\ (2008). If you can compute the first 10 eigenvalues of $\Delta|_{G\text{-inv}}$ and confirm they correspond to $p_2, e_3$, that is strong evidence for Conjecture~\ref{conj:strong}. \end{enumerate} % ------------------------------------------------------------------ \section{Reading List} % ------------------------------------------------------------------ \textbf{Essential (read first):} \begin{enumerate} \item S.~K.~Donaldson, ``Scalar curvature and projective embeddings I,'' J.\ Diff.\ Geom.\ \textbf{59} (2001), 479--522. \textit{[Defines the T-operator and balanced metrics; proves convergence to CY.]} \item S.~Zelditch, ``Szeg\H{o} kernels and a theorem of Tian,'' Int.\ Math.\ Res.\ Not.\ \textbf{6} (1998), 317--331. \textit{[The foundational TYZ expansion paper.]} \item Z.~Lu, ``On the lower order terms of the asymptotic expansion of Tian-Yau-Zelditch,'' Am.\ J.\ Math.\ \textbf{122} (2000), 235--273. \textit{[Explicit computation of $a_1, a_2$ in the TYZ expansion.]} \end{enumerate} \textbf{Important context:} \begin{enumerate} \setcounter{enumi}{3} \item J.~Fine, ``Calabi flow and projective embeddings,'' J.\ Diff.\ Geom.\ \textbf{93} (2013), 55--99. \textit{[Connects Bergman kernels to the balanced metric iteration rigorously.]} \item X.~Wang, ``Canonical metrics on stable vector bundles,'' Commun.\ Anal.\ Geom.\ \textbf{13} (2005), 253--285. \textit{[Good background on balanced metrics and convergence rates.]} \item V.~Braun, T.~Brelidze, M.~R.~Douglas, B.~A.~Ovrut, ``Eigenvalues and eigenfunctions of the scalar Laplace operator on Calabi-Yau manifolds,'' JHEP \textbf{2008} (2008), 120. \textit{[Numerically computes Laplacian eigenvalues on the Fermat quintic. Gives you the spectral data you need for Lemma 5.]} \item R.~P.~Stanley, ``Invariants of Finite Groups and their Applications to Combinatorics,'' Bull.\ Amer.\ Math.\ Soc.\ \textbf{1} (1979), 475--511. \textit{[Invariant ring theory needed for Lemma 1.]} \end{enumerate} \textbf{Background geometry:} \begin{enumerate} \setcounter{enumi}{7} \item P.\ Griffiths and J.\ Harris, \textit{Principles of Algebraic Geometry}, Wiley (1978). Chapters 1.5 (second fundamental form) and 3.3 (Calabi-Yau). \item D.\ Huybrechts, \textit{Complex Geometry: An Introduction}, Springer (2005). Chapter 3 (K\"ahler manifolds) and Chapter 5 (Calabi-Yau). \end{enumerate} % ------------------------------------------------------------------ \section{Summary of What This Paper Would Be} % ------------------------------------------------------------------ \begin{center} \renewcommand{\arraystretch}{1.5} \begin{tabular}{p{3.5cm}|p{10cm}} \hline \textbf{What it is NOT} & An ML paper. Symbolic regression appears only as a computational tool in Section 4 to identify the rational formula. \\ \hline \textbf{What it IS} & A mathematical physics paper on the $G$-invariant structure of the balanced metric on the Dwork quintic family. \\ \hline \textbf{Main object} & $R_k(z) = \log(\det g_k / \det g_{\mathrm{FS}})$, the balanced metric volume-form ratio. \\ \hline \textbf{Main result} & This function is $G$-invariant (theorem), lives in a 4-generator ring (theorem), and is dominated by its projection onto $\mathrm{span}(p_2, e_3)$ (conjecture + evidence, or theorem if TYZ works). \\ \hline \textbf{Evidence type} & Algebraic proof (invariant ring) + analytic computation (scalar curvature) + numerical support (ablation, k-stability). \\ \hline \textbf{Target} & \textit{Letters in Mathematical Physics} (conjecture + evidence) or \textit{Communications in Mathematical Physics} (with theorem). \\ \hline \textbf{Timeline} & 2--3 months to a Letters in Math Phys submission; 6--12 months to a CMP-level result. \\ \hline \end{tabular} \end{center} \end{document}