Papers

This is the third of three related preprints. We consider here which groups $\pi$ and $PD_3$-complexes $Y$ are realised by $PD_4$-pairs $(X,Y)$ with $X$ aspherical and $\pi_1X\cong\pi$, and show that such a pair may be assembled from $(D^4,S^3)$ and $PD_4$-pairs of groups, by adding 1-handles and mapping cylinders of $\mathbb{Z}\pi_1N$-homology equivalences over boundary components $N$, if and only if $H^2(\pi;\mathbb{Z}\pi)=0$. (This includes all such pairs with $\pi_1$-injective boundary components and all with $\pi$ a free group, but none with $c.d.\pi=2$. Adding a 1-handle includes connected sum.) If there is a finite 2-dimensional $K(\pi,1)$-complex then $\pi$ is realisable by some such pair $(X,Y)$, but there are no obvious building blocks analogous to $PD_4$-pairs of groups, except for when $\pi$ is a $PD_2$-group.

We consider the problem of when it is best to sell a risky asset in the distribution builder framework under the consideration of potential ruin. This approach allows investors to express their preferences directly as a desired target distribution without first specifying a risk aversion or utility function. Mathematically, the problem is closely related to the Skorokhod embedding problem of attaining a given distribution by stopping a diffusion process. We work in a general framework of one-dimensional diffusion processes and extend existing results to include the possibility of ruin. We first provide a full characterization of the set of distributions that can be attained before ruin occurs. Then, we formulate two optimization problems that tackle the issue of what to do if the originally specified distribution is either not attainable or not optimal: finding the attainable distribution closest to an unattainable distribution and selecting an optimal attainable distribution under first-order stochastic dominance constraints if the originally specified distribution is attainable, but not optimal. We provide sufficient conditions for the existence and uniqueness of solutions to these constrained convex optimization problems in an extended \(f\)-divergence framework, provide an analytic characterization of the solutions, and present numerical examples.

We study the orbital stability of solitary waves to the generalized Benjamin–Bona–Mahony equation $$ u_t+u_x+\kappa\frac{p+1}{2}(u^p)_x-u_{txx}=0, \qquad (t,x)\in\mathbb R^+\times\mathbb R, $$ where $\kappa\in\mathbb R\setminus\{0\}$ and $p\geq2$ is an integer. This equation admits solitary waves of the form $$ u(t,x)=\phi_c(x-ct). $$ At the negative critical endpoint speed $$ c=c_p^-= \frac{p-1}{2(p+1)} \left(1-\sqrt{\frac{p+3}{2}}\right)<0, $$ we prove that the corresponding solitary wave is orbitally unstable whenever either $\kappa<0$, or $\kappa>0$ and $p$ is even. The proof combines a codimension-two coercivity estimate, two-parameter modulation, and a corrected localized virial functional adapted to the degeneracy of the momentum slope at the endpoint.

Diffusion models are widely used as priors for linear inverse problems, yet endpoint quality does not reveal when measurement information enters reverse denoising or how it is allocated across signal directions. We study this process through the smoothed likelihood force, the difference between exact posterior and prior scores at each noise level. For a fixed measurement, its expected squared norm gives both posterior–prior relative-entropy dissipation and reverse-path relative-entropy growth. Averaging over measurements yields an information–minimum mean-square error (I-MMSE) identity linking information gain to denoising-error reduction. Under finite second moments, the force energy and its ratio to prior-score energy decay quadratically in the noising kernel's signal coefficient at high noise. Solvable models show that conditioning removes class separation already explained by the measurement, reduces a uniform index entropy over \(n\) empirical samples from \(\log n\) to \(H(I\mid r)\), and makes assimilation depend on operator–prior alignment even for identical singular values. Experiments in models with tractable posteriors evaluate these predictions. In a separate illustration with a frozen FFHQ model, masks sharing the same spectrum yield different prior-normalized null-space trajectory statistics.

A long-standing open question in quasi-Newton optimization asks whether the classical Davidon–Fletcher–Powell (DFP) method converges globally on uniformly convex objectives when all accepted steps satisfy the standard weak Wolfe conditions. We show that the answer is no, even under the standard strong Wolfe conditions. Fix $0<c_1<2/3$ and $2/3\le c_2<1$. We construct a function $f\in C^2(\mathbb{R}^2)$ such that $\frac{1}{2}I\preceq\nabla^2 f(x)\preceq\frac{3}{2}I$ for all $x\in\mathbb{R}^2$. We also choose a fixed positive definite initial inverse Hessian approximation and a sequence of positive step lengths. The classical DFP iteration is well defined, and all accepted steps satisfy the standard strong Wolfe conditions, but $|\nabla f(x_k)|$ converges to a positive constant. The global Hessian condition number is at most three. The construction uses an alternating two-step DFP sequence near a one-dimensional invariant center manifold. Along this sequence, the smaller eigenvalue of the inverse Hessian approximation tends to zero. The changes in the gradient norm between cycle starts are summable, but the total rotation of the associated eigenvectors is unbounded. The accumulation points of the DFP sequence form a circle. A uniform separation bound allows us to interpolate the prescribed function values and gradients. We add smooth functions with pairwise disjoint supports to a quadratic and keep the global Hessian bounds. An affine change of variables gives an identity-initialized example with problem-dependent Hessian bounds. An orthogonal direct sum extends the result to every dimension $n\ge 2$.

Let $M$ be a countable transitive model of a sufficiently large finite fragment of ZFC, and let $E^{\mathbb B}_{M}$ be equivalence of random generics over $M$. Smythe proved that $E^{\mathbb B}_{M}$ is not Borel-reducible to the orbit relation of a free action of any countable group belonging to $M$, and asked whether the restriction on the target group can be removed. We prove that if $A$ is a positive-measure Borel set of random reals over $M$ and $F$ is an essentially free countable Borel equivalence relation, then every Borel homomorphism from $E^{\mathbb B}_{M}\upharpoonright A$ to $F$ maps a conull subset of $A$ into a single $F$-class. Hence no positive-measure restriction of $E^{\mathbb B}_{M}$ is essentially free, or even weakly Borel-reducible to an essentially free relation. The proof combines Thomas's consequence of Popa cocycle superrigidity with a perfect family of $2$-marked groups and a Fubini argument using two successive random reals.

The randomized extended Kaczmarz method is an effective iterative framework for solving large-scale inconsistent linear systems. In this paper, we extend this framework to third-order inconsistent tensor linear systems under the t-product and propose the Tensor Greedy Double Block Extended Kaczmarz (TGDBEK) method. At each iteration, TGDBEK dynamically constructs active blocks of row and column slices via a residual-based greedy selection strategy, prioritizing the slices associated with the largest residual norms. Unlike existing tensor block Kaczmarz variants that rely on static, predefined partitions –the tensor randomized extended block Kaczmarz (TREBK) method, its greedy counterpart (TREGBK), and the tensor randomized extended average block Kaczmarz (TREABK) method – TGDBEK adapts the active block sizes dynamically at each step using a single intuitive threshold parameter $\eta$. We establish the theoretical linear convergence of TGDBEK to the unique minimum-norm least-squares solution $\mathcal{A}^\dagger * \mathcal{B}$. Extensive numerical benchmarks on synthetic dense and sparse tensor systems, as well as multidimensional multichannel color and 3D volumetric MRI image deblurring problems, demonstrate that TGDBEK substantially outperforms state-of-the-art tensor Kaczmarz solvers in both iteration count and CPU running time.

We show that the inviscid shallow water equations with gravity and surface tension are locally well-posed in dimensions one and two. In contrast to previous works by Benzoni-Gavage, Danchin, and Descombes, we work only in the real-valued phase space and make use of the symmetry of the system. Our alternative proof for local well-posedness sheds some light on the symmetric structure of the shallow water system and can be used for further study, including designing numerical scheme, dispersion estimate, etc.

Every topologically mixing shift of finite type has a unique measure of maximal entropy. It follows from Ornstein theory that every invariant measure with entropy close to maximal must be close to the MME in the $\bar{d}$-metric. We prove a quantitative version of this, strengthening earlier results of Kadyrov that established effective uniqueness in the Wasserstein metric.

We study the superdiffusive behavior of random walks on the randomly oriented Manhattan lattice, i.e., the $d$-dimensional integer lattice $\mathbb{Z}^d$ where each axis-aligned line is independently assigned a random direction (forward or backward) with equal probability. The walker takes nearest-neighbor steps, choosing an axis randomly and moving along the assigned direction of that axis's line, with equal probabilities for each axis. We show that, in the critical dimension $d=3$, the diffusion coefficient of the random walk diverges in the Tauberian sense as $\sqrt{\log t}$ with a multiplicative correction $(\log\log t)^{\pm(2+\varepsilon)}$ as time $t\to\infty$. This gives an answer to a conjecture by Ledger, Tóth and Valkó (2018).

Physics-Informed Neural Networks approximate the solution of boundary value problems by minimizing the strong-form residual, their accuracy is often limited near corner and crack singularities, sharp layers and localized features and the enforcement of Neumann boundary conditions remains challenging. In hp-Variational Physics-Informed Neural Networks, the problem is formulated in a Petrov-Galerkin weak form, reducing the differentiation order applied to the neural network, while boundary constraints are imposed through penalty terms. The variational residual is evaluated using fixed Gauss quadrature points and the strong-form residual on a fixed collocation set, which may fail to capture regions of rapid variation. To address this issue, we integrate hp-Variational Physics-Informed Neural Networks with Residual-based Adaptive Refinement. During training, the strong-form residual is computed on a dense candidate pool and points with the largest residuals are added adaptively to the collocation set. Since the weak formulation is defined in $H^{1}(\Omega)$, it avoids the $H^{2}(\Omega)$ regularity requirement that fails on cracked and non-convex domains. The proposed method is validated on four two-dimensional elliptic problems and consistently achieves substantial reductions in relative $L^{2}$, absolute $L^{\infty}$ and relative $H^{1}$ semi-norm errors through adaptive residual-driven point placement.

For every fractional order strictly between zero and one and every finite interval, there is a bounded representing function whose Riemann–Liouville fractional integral satisfies the weighted endpoint condition proposed by Vainikko but fails the absolute Marchaud condition almost everywhere. One bounded example therefore disproves the necessity implication in Vainikko's conjectured characterization simultaneously for every exponent strictly larger than one, including the essentially bounded endpoint. A different sine-series construction gives, at each exponent strictly between one and infinity, failure at every interior point for one representative. For the bounded example, signed inversion remains valid for every finite exponent greater than or equal to one. The converse implication in Vainikko's conjecture is not decided here.

Chen, Mohar and Wu [J. Combin. Theory Ser. B 161 (2023)] conjectured that every graph of orientable genus $g$ has proper orientation number $O(\sqrt g)$. We prove this conjecture. Our main result extends the potential-outdegree method from independent color classes to three parts of bounded internal degree. Let $k,D\in\mathbb Z_{\ge0}$. If $G$ has maximum average degree at most $2k$ and $V(G)=V_1\cup V_2\cup V_3$, where each $V_i$ induces a subgraph of maximum degree at most $D$, then $$\pchi(G)\le k+9D+6+\floor{\frac32\ceil{\frac{2D+2}{5}}}\le k+\floor{\frac{48D}{5}}+8.$$ When the partition is given, we construct such an orientation in polynomial time. For every $D\ge1$, we also construct a graph with a partition into three parts of internal maximum degree at most $D$ for which $\pchi(G)-\ceil{\mad(G)/2}\ge6D+5$; hence the linear dependence on $D$ is unavoidable. Finally, Woodall's defective-coloring theorem yields $\pchi(G)=O(\sqrt{\gamma+1})$ for every graph of Euler genus $\gamma$.

Consider testing a finite composite null $\cP={P_1,\ldots,P_L}$ against a simple alternative $Q$ using $n$ i.i.d. observations. Let $\ell_n$ denote the supremum of the expected log e-value over e-variables that are exact under every $P_i$ and pivotal across the nulls. Zhang, Ramdas, and Wang (2024) showed that $\ell_n$ is superadditive and asked whether $\ell_n/n$ converges to the upper bound $\min_i D(Q|P_i)$ under their joint atomlessness and absolute continuity assumptions. We prove that it does. The proof draws on techniques developed by Zhang, Ramdas, and Wang (2024) and by Farooq, Fritz, Haapasalo, and Tomamichel (2024).

Designing a high-order structure-preserving discretization is a constrained mathematical search: conservation, a positive discrete inner product, physical spectral behavior, boundary accuracy, bandwidth, and partial differential equation (PDE) error must hold simultaneously. We test whether large language models (LLMs) can help while remaining non-authoritative. The motivating MOLE implementation of the Corbino-Castillo staggered operators satisfies a general discrete Gauss identity and conserves exactly, yet its order-six and order-eight Dirichlet blocks develop four non-real boundary-localized modes. An endpoint-supported positive-definite identity would instead force a real non-positive spectrum, so the search changes the closure and norm architecture. An LLM proposes only a typed construction program; a deterministic linear-program compiler generates coefficients; an independent verifier tests algebra, positivity, physical modes, conditioning, and manufactured PDEs; and coupled rational reconstruction provides exact certificates. Across 1,200 solver evaluations, an externally fixed verifier accepted 55.0% of full-metric-feedback proposals and 53.3% of illumination-archive proposals, versus 13.3% for uniform random search. Four leading LLM-originated programs were reconstructed exactly. The strongest order-six-interior, order-four-boundary candidate lowers the prior positive-diagonal spectral-radius constant by 25.4% and its PDE error 62.7-fold. Its certified heat-equation energy is contractive, whereas the order-six reference exhibits 6.6% transient growth; both have the same RK4 stability limit. The LLM proposes structural hypotheses; deterministic mathematics determines validity.

Neural combinatorial optimization solvers generate many candidate solutions per instance and report the best one found, using the same sample budget for every instance regardless of difficulty. A companion study showed that reallocating a fixed budget toward harder instances can improve solution quality, but that the standard way of measuring this improvement is biased: deciding an allocation and evaluating it on the same data can manufacture an apparent gain even when none exists. This left open what property of a workload determines whether reallocation is worth doing, and whether a policy that spends part of the budget to decide how to allocate the rest still pays once that cost is counted. This paper answers both questions through pre-registered confirmatory experiments – analysis and verdict criteria fixed before data collection – across three independently trained solvers and two ways of constructing harder workloads on the traveling salesman problem. Within the workloads we study, the deciding property is how varied the instances within a workload are in difficulty, not how difficult the workload is on average: a uniformly easy or uniformly hard workload offers little room for reallocation, while a mixed workload offers substantial room. A budget-aware policy that pays for its own information about instance difficulty recovers most, though not all, of the improvement available when that information is assumed free. Every experiment was independently recomputed from its written specification, and every correction to an earlier version – including two that weakened the paper's own claims – is reported with the direction it moved the conclusion. The paper offers a specific empirical answer and a template for verifying that answer is not an artifact of how it was measured.

We study zeros transported by the heat semigroup from stationary solutions of the sparse operators $L_n=c_{n,0}\D^n+c_{0,n-1}x^{n-1}$. The starting point is the exact cubic boundary obtained by Hernández-del-Valle and Guerra-Polania~\cite{HerGuerra}. We show that the hypergeometric function used there is the residue-$1$ component of a naturally graded stationary solution space, and that heat transport preserves this grading after the scaling $x=t^ny$. This identifies the cubic boundary as the transported zero of a specific stationary mode. We then study $L_4=c_{4,0}\D^4+c_{0,3}x^3$. Its $6\times6$ boundary-jet determinant leads to the universal scaling \[ F(t)=\rho\left(\frac t2\right)^4 \Phi\!\left(\rho^2\left(\frac t2\right)^7\right), \qquad \rho=\frac{c_{0,3}}{c_{4,0}}. \] The resulting nonlinear equation for $\Phi$ is singular at the origin. After an explicit desingularization, its initial compatibility condition factors as \[ (5\Phi'(0)-1)(6\Phi'(0)-1)=0. \] Consequently there are exactly two normalized graded analytic boundary germs. We place the desingularized equation in Briot–Bouquet form, prove local existence and uniqueness of both branches, and give a triangular recursion for all their Taylor coefficients. Finally, we prove that the two branches are generated respectively by the residue-$1$ and residue-$3$ stationary solutions. The quadratic and cubic cases provide the established baseline; the fourth-order problem is the first point at which distinct stationary modes produce distinct analytic branches. We do not claim a general branch-count theorem.

Let $A$ be a stochastic $n\times n$ matrix with $\operatorname{rank}A\leq k$, where $1\leq k\leq n$. Write $n=qk+s$, where $0\leq s<k$. The author conjectured in 2018 that $\operatorname{per}A\leq (q!/q^q)^{k-s}((q+1)!/(q+1)^{q+1})^s$, with equality if and only if $A=P(J_q^{\oplus(k-s)}\oplus J_{q+1}^{\oplus s})Q$, where $P,Q$ are permutation matrices and $J_t$ is the $t\times t$ matrix with every entry $1/t$. We prove this conjecture in full.

A quantum feedback loop that returns information from a forward simulation to an earlier internal time induces a completely positive trace-preserving map on a message register; iterating that map eventually destroys its ability to carry entanglement. The entanglement-breaking index N = n_EB(Phi), the round at which this happens, is studied along two axes, interaction and bath. On the interaction axis, strict contraction onto a full-rank fixed point forces N finite in every dimension, via an explicit separable ball around the limiting Choi state, and across 3997 sampled channels index and stability gap are one clock, N [1 - rho(A)] in [0.60, 1.50]. On the bath axis, a thermal ancilla at polarisation p = tanh(hbar omega / 2 k_B T) makes every channel quantity an exact quadratic in p, and one-sided certificates (no eigensolver) establish N = 3 at the reference circuit at every temperature. The infinite-temperature index is an architecture floor with closed form and exact measure (46.427% of circuits break entanglement in one round at infinite temperature). For exchange collisions (generalised amplitude damping) the index has a closed form and is monotone in p; for the loop the floor is not a bound: certified circuits dip below it in interior temperature windows, at a rate with Gaussian cutoff in epsilon = 1 - ||A_0||_2 (marginal exponent 2.02 [1.90, 2.16], 4.1 million circuits), and shallow-floor valleys survive to epsilon = 0.125. On a certified coupling-temperature surface, temperature moves the index by at most 3/2, the interaction by 1213/3. Applied to quantum-cognition proposals, a per-bath polarisation table at 310 K shows nuclear spins maximally hot and intramolecular vibrations frozen, and a four-input checklist replaces decoherence times by round budgets; Fisher's nuclear-spin proposal specifies its bath and storage coherence, not its processing interaction or rate.

For a channel $\Phi$ on $M_d$ written in the generalised Bloch parameterisation $r \mapsto Ar+c$, we prove that entanglement breaking implies $\|A\|_*^2 + \frac{d(d-1)}{2}\,|c|^2 \;\le\; (d-1)^2 ,$ where $\|\cdot\|_*$ denotes the nuclear norm. Read on the Choi state of the channel, the inequality is an instance of the covariance-matrix criterion of Gittsovich et al.; what is added here is an elementary proof in Holevo form and a complete description of the equality cases. Apart from the constant channel onto a pure state, which shows that the coefficient $d(d-1)/2$ cannot be increased, equality holds exactly for the measure-and-prepare channels with rank-one effects $\alpha_k|e_k\rangle\langle e_k|$ and pure outputs $|\psi_k\rangle\langle\psi_k|$ whose overlaps are an affine image of those of the measurement directions, $|\langle\psi_j|\psi_k\rangle|^2 = P + (1-P)\,(d|\langle e_j|e_k\rangle|^2 - 1)/(d-1)$ for a common $P = \mathrm{Tr}\,\Phi(\mathbf{1}/d)^2$. These include the completely dephasing channel, the re-preparation channel of every rank-one POVM, a one-parameter family interpolating between dephasing and the constant channel, and non-unital non-projective examples already at $d = 2$. The qubit case reads $\|A\|_*^2+|c|^2\le1$ and extends to full rank a rank-two condition of Ruskai. The proof sharpens Ruskai's bound $\|A\|_*\le1$ by retaining the POVM completeness relation $\sum_k f_k=0$, which her argument discards; this recentres $A$ from a second moment into a cross-covariance, and is the entire content of the improvement.

For every integer $m\geq 3$, we construct a smooth Riemannian metric $g$ on a manifold diffeomorphic to $S^m$ such that $\text{sec}_g\geq 0$, every sectional curvature is strictly positive away from a closed nowhere-dense set, and the point locus on which some sectional curvature vanishes has Hausdorff dimension exactly $\frac{1}{2}$. The metric is induced on the boundary of a smooth convex body in $\mathbb{R}^{m+1}$. The essential local model is a convex graph whose Hessian has a one-dimensional kernel precisely on a Cantor subset of one coordinate axis; a regularized maximum then inserts this graph into a round sphere without introducing additional degenerate points. The main content of this paper is generated by ChatGPT 5.6 and verified by the author.

Fractional Laplacians encode anomalous scaling and covariant nonlocal response in quantum matter and holographic boundary systems. We place the corresponding scalar spectral theory on a dynamical closed worldsheet. A metric-dependent intertwiner maps the fractional quadratic form to that of $D$ free bosons, while its zeta Jacobian changes the matter determinant weight from $D$ to $sD$. The prime-determinant transformation includes the zero-mode area factor and determines both the moduli density and the Weyl coefficient $c_{\rm det}=sD$. Weyl cancellation therefore selects $D_\ast=26/s$. On the integer critical locus $1\leq D\leq26$, $N=26-D$ spectator bosons turn the residual determinant into a local critical CFT. Its fixed-area genus-$g$ vacuum measure is that of 26 free bosons, its torus trace is modular invariant, its plumbing channels are state resolved, and its BRST charge is nilpotent. The $D$ distinguished coordinates remain spectral composites; their pullback yields the Koba–Nielsen amplitudes and massless vertex conditions. Target backgrounds coupled to this distinguished sector retain the ordinary one-loop tensor beta functions, with dilaton deficit $sD-26$ and the corresponding string-frame action. Worldsheet gravity thus turns the fractional exponent into a quantum-geometric consistency parameter.

For a tree $T$, let $[T]$ be the set of isomorphism classes of trees that are mutually topological minors of $T$. Bruno and Szeptycki proved that every locally finite tree satisfies $|[T]|\in\{1,2^{\aleph_0}\}$ and that every tree with a ray containing infinitely many vertices of degree at least $3$ has at least $2^{\aleph_0}$ topological twins. Hence equality holds in the latter case when the tree is countable. We treat the complementary countable case, without any bound on the degrees. For every such tree $T$, we prove that $|[T]|\in\{1,\aleph_0,2^{\aleph_0}\}$ and give a well-founded recursive description based on Schmidt rank. The proof uses a finite canonical subtree extracted from the subtree generated by the branching vertices and a counting theorem for countable multisets over a well-quasi-order. The same trichotomy follows for every countable tree and hence for every tree all of whose vertices have countable degree.

We prove a law of fractional logarithm for the largest eigenvalue along a northwest-nested path of complex sample covariance matrices from one infinite array. The entries are independent and centered, with unit variance, vanishing complex second moment, and uniformly bounded moments of every fixed order. The row dimension is nondecreasing, has bounded increments, and has a positive limiting aspect ratio. After finite-size edge centering and scaling, the almost-sure limsup on the $(\log N)^{2/3}$ scale is $(1/4)^{2/3}$, and the liminf on the $(\log N)^{1/3}$ scale is $-4^{1/3}$. The corresponding cluster sets in $\mathbb{R}$ are $[0,(1/4)^{2/3}]$ and $[-4^{1/3},\infty)$. The proof compares the Laplace transform of a single smoothed count over a growing grid of full nested matrices with its Gaussian counterpart. Gaussian count concentration gives block occurrences with probability tending to one. Dyadic tail bounds yield the endpoints, and deterministic interpolation gives the cluster sets.

A cycle of nine pitch classes in twelve-tone equal temperament can be arranged so that its nine consecutive three-note windows realize, exactly once each, every trichord type containing a semitone – types taken up to transposition but not inversion. Such cycles exist in abundance, and generalize to n-tone equal temperament, where the windows realize the n-3 semitone-containing types. In every one of the 1764 such cycles in twelve-tone equal temperament, the chromatic trichord 012 appears scalarly, as a direct chromatic run, and never in a broken "zigzag" contour such as 1-0-2. Why is the zigzag missing? We show that the zigzag presentation behaves as a conserved topological charge of an underlying interval-flow network: it occurs if and only if 3 divides n and n is neither 6 nor 12. The forward obstruction is a Z/3-valued conservation law, proved for all n by a weight-function argument; the two exceptional temperaments – one of them the common Western tuning – are excluded by finite certificates; and existence for every other multiple of 3 follows from an explicit construction. We also quantify rarity, and give exact censuses well beyond the reach of direct search. The classification is machine verified in Lean 4 in both directions. The paper is written for two audiences. Musical sections require no more than pitch-class set vocabulary; the mathematical development is self-contained and elementary, yet fully rigorous.

Discrete energy-level structures in Coulomb-like localized bound states are conventionally obtained by direct diagonalization of the Schrödinger equation or by semiclassical approximations that fail at low energies. Here we develop a geometric phase-space framework that recovers the exact discrete energy levels and shell degeneracy of three-dimensional Coulomb bound states from classical symplectic geometry and global boundary constraints, bypassing explicit operator diagonalization. Using Kustaanheimo–Stiefel regularization, we map negative-energy Kepler orbits to four-dimensional isotropic harmonic oscillators. The group composition law of the symplectic flow determines an integral kernel built from the classical two-point action and Van Vleck amplitude, whose infinitesimal generator satisfies an exact linear evolution equation—a property we term quadratic closure. Imposing three geometric constraints—decay at infinity, regularity at the origin, and fiber invariance under the Hopf fibration—we recover the exact $E_n/E_1=1/n^2$ scaling and $g_n=n^2$ degeneracy. The phase scale $\alpha$ sets the action unit; all spectral structural features are independent of its value. This framework offers a symmetry-transparent route to spectral data: for quadratic Hamiltonians, the classical kernel incurs zero semiclassical error even at ground-state energies, and the geometric constraints directly identify symmetry-protected degeneracies without diagonalization. We apply the method to shallow impurity states, moiré superlattices, and semiclassical transport, demonstrating how geometric screening rules simplify spectral analysis in condensed-matter environments.

We extend the discrete gauge rigidity of time-scaled intertwining cocycles to operator networks over a connected manifold M and classify the holonomy that the continuous setting admits. For dissipative semigroups $S_x(t)=e^{-tA_x}$ linked by transport operators $K_\gamma S_{\gamma(0)}(t)=S_{\gamma(1)}(\lambda(\gamma)t)K_\gamma$ along paths $\gamma$, we prove that the scaling field carries no monodromy: $\lambda(x,y)=\tau(x)/\tau(y)$ for a continuous positive function $\tau$, unique up to a multiplicative constant, with multiplicativity forced by the intertwining relation rather than assumed. For regular cocycles the transport defines a bounded operator connection on a Hilbert bundle over M; the holonomy of a closed loop is confined to the commutant of the generator and decomposes into spectral sectors, on each of which it is adiabatic (Berry-Wilczek-Zee) transport twisted by an independent commutant-valued sector potential; for flat unitary cocycles the holonomy is classified by unitary representations of $\pi_1(M)$ in the commutant. An explicit rotating network over $S^1$ exhibits holonomy equal to the parity operator while the Berry connection one-form of every level vanishes identically: the monodromy is carried by $\mathbb{Z}_2$ orientation classes of Moebius eigenline bundles.

This paper is a corrected English translation of an original Japanese work introducing "power geometry," a research programme centred on power transformations, deformations and dualities of functions in the sense of optimisation theory. Its two main motivations are the origin of probability distributions (via maximum-likelihood principles, rational-function parameters and ODEs linked to information geometry, the Pearson system and Schwarzian equations) and the construction of matrix dynamical systems that compute eigenvalues, singular values and principal or minor subspaces. Building on Brockett's double-bracket flows, the author extends these systems to rectangular matrices, flag manifolds and difference-of-convex potentials, and shows how a continuous power parameter can interpolate between largest- and smallest-eigenvalue problems. Dual potentials, topological distinctions between principal and minor subspace flows, and links via matrix Schwarzian–Riccati equations are examined, with a view toward design principles for classical and quantum information-processing algorithms.

We study relative entropy under gauge reduction in Whitney-discretized electromagnetism. The discrete Hodge decomposition separates exact gauge directions from harmonic and coexact physical degrees of freedom. For Gaussian distributions on the auxiliary potential space, relative entropy splits exactly into the relative entropy of the physical marginals and a nonnegative conditional gauge contribution. We express this contribution through Schur complements, conditional means, and physical–gauge correlations, and characterize equality between auxiliary and physical relative entropies. The physical divergence is also shown to be the minimum over all Gaussian auxiliary extensions. At the infinitesimal level, Fisher information admits an analogous decomposition. A two-variable model illustrates how statistically distinct auxiliary potentials may represent identical physical statistics. The construction provides an information-geometric interpretation of discrete gauge reduction without assigning physical significance to the eliminated gauge variables.

We introduce the notion of universal quilt for toric CY3 singularities and construct examples when the universal quilts are themselves smooth Calabi-Yau threefolds. Such quilts are designed to contain all of the crepant resolutions of the given singularity. To obtain universal quilts we implement an algorithm that computes all triangulations for any given planar polygon. We also introduce the notion of quilt stack, to address the question of finding the most likely smooth configuration preceding a given singularity.