Peer Revenue Equity in science

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Mathematics

math Watch Mathematics (sign in to use)

Thursday 5 March

Numerical Analysis

Everything is Vecchia: Unifying low-rank and sparse inverse Cholesky approximations

The partial pivoted Cholesky approximation accurately represents matrices that are close to being low-rank. Meanwhile, the Vecchia approximation accurately represents matrices with inverse Cholesky factors that are close to being sparse. What happens if a partial Cholesky approximation is combined with a Vecchia approximation of the residual? This paper shows how the sum is exactly a Vecchia approximation of the original matrix with an augmented sparsity pattern. Thus, Vecchia approximations subsume a class of existing matrix approximations and have broad applicability.
Read the full abstract

Combinatorics

Color $2$-switches and neighborhood $\lambda$-balanced graphs with $k$ colors

This paper examines vertex colorings of graphs with constraints on the distribution of colors in vertex neighborhoods. We introduce color 2-switches and color degree matrices. The color degree matrix of a $k$-colored graph is an analog of the degree sequence, while a color 2-switch provides a way to transform a $k$-colored graph to another such graph while maintaining the color of each vertex and the multiset of colors in each vertex neighborhood. We prove that two $k$-colored graphs have the same color degree matrix if and only if one can be obtained from the other by a sequence of color 2-switches. In related work, we generalize neighborhood balanced colorings by allowing for $k$ colors (instead of two) and more flexibility on the number of vertices of each color in a neighborhood. We introduce three classes of $k$-colored, $\lambda$-balanced graphs, in which any two color classes in a vertex neighborhood differ in size by at most $\lambda$. These classes are distinguished by whether the balancing condition is imposed on the open neighborhood $N(v)$, the closed neighborhood $N[v]$, or allowed to vary by vertex. For each class, the minimum $\lambda$ for which a graph admits a balanced coloring defines its $\lambda$-balance number. We prove general results about these classes and their $\lambda$-balance numbers. For $k = 2$, we introduce a fourth class, parity balanced graphs, in which the number of vertices of each color are equal in open neighborhoods for even-degree vertices and in closed neighborhoods for odd-degree vertices. Additionally, we focus on the important case where $k=2$ and $\lambda \le 1$ and introduce the technique of red-blue removals. We provide separating examples between these four classes and prove balance number results for paths, cycles, wheels, trees, caterpillars, and complete multipartite graphs, and a counting result for caterpillars.
Read the full abstract

All authors Karen L. Collins, Jonelle Hook, Cayla McBee, Ann N. Trenk

Statistics Theory · Cross-listed from stat.ME

Random Dot Product Graphs as Dynamical Systems: Limitations and Opportunities

Can we learn the differential equations governing the evolution of a temporal network? We investigate this within Random Dot Product Graphs (RDPGs), where each network snapshot is generated from latent positions evolving under unknown dynamics. We identify three fundamental obstructions: gauge freedom from rotational ambiguity in latent positions, realizability constraints from the manifold structure of the probability matrix, and trajectory recovery artifacts from spectral embedding. We develop a geometric framework based on principal fiber bundles that formalizes these obstructions. We characterize invisible dynamics as exactly the skew-symmetric generators, and show the realizable tangent space has dimension $nd - d(d-1)/2$. An holonomy dichotomy emerges: polynomial dynamics have commuting generators, stationary eigenvectors, and trivial holonomy, making gauge alignment purely statistical; Laplacian dynamics satisfy a non-commutativity criterion producing nontrivial holonomy, with curvature weighted by $1/(\lambda_\iota + \lambda_\gamma)$ linking gauge sensitivity to the spectral gap. In $d=2$ this yields full restricted holonomy $\mathrm{SO}(2)$; for $d \ge 3$ generic full $\mathrm{SO}(d)$ remains conjectural. Cram'er–Rao lower bounds reveal that the same spectral gap controlling curvature and injectivity simultaneously controls Fisher information, so geometric and statistical difficulty are inextricable. We prove an identifiability principle: symmetric dynamics cannot absorb skew-symmetric gauge contamination, so dynamics structure can resolve gauge ambiguity. We demonstrate this constructively with anchor-based alignment and a UDE pipeline recovering vector fields from noisy graph sequences. Yet finite-sample interactions between noise, gauge, and dynamics expressiveness remain beyond the asymptotic theory. We frame this gap as an open challenge.
Read the full abstract

Combinatorics

Pseudo-orientable ribbon graphs: Matrix–Quasi-tree Theorem and log-concavity

One of the most important classes of even $\Delta$-matroids arises from orientable ribbon graphs, which play a role analogous to that of graphic matroids in matroid theory. Motivated by a natural correspondence between strong $\Delta$-matroids and even $\Delta$-matroids due to Geelen and Murota, we characterize the class of strong $\Delta$-matroids that correspond to orientable ribbon-graphic $\Delta$-matroids. These are precisely the $\Delta$-matroids associated with what we call pseudo-orientable ribbon graphs. Moreover, we present a geometric construction that transforms a pseudo-orientable ribbon graph into an orientable ribbon graph, thereby realizing this correspondence. As consequences, we obtain the Matrix–Quasi-tree Theorem, the Hurwitz stability of quasi-tree generating polynomials, and a log-concavity result for the sequence counting quasi-trees of size $2i-1$ or $2i$ for pseudo-orientable ribbon graphs. To establish the log-concavity, we generalize Stanley's log-concavity theorem for regular matroids to regular $\Delta$-matroids. Finally, we exhibit an infinite family of non-pseudo-orientable ribbon graphs that fail to satisfy the Matrix–Quasi-tree theorem and Hurwitz stability.
Read the full abstract

Numerical Analysis · Cross-listed from cs.CE

Autonomous Algorithm Discovery for Ptychography via Evolutionary LLM Reasoning

Ptychography is a computational imaging technique widely used for high-resolution materials characterization, but high-quality reconstructions often require the use of regularization functions that largely remain manually designed. We introduce Ptychi-Evolve, an autonomous framework that uses large language models (LLMs) to discover and evolve novel regularization algorithms. The framework combines LLM-driven code generation with evolutionary mechanisms, including semantically-guided crossover and mutation. Experiments on three challenging datasets (X-ray integrated circuits, low-dose electron microscopy of apoferritin, and multislice imaging with crosstalk artifacts) demonstrate that discovered regularizers outperform conventional reconstructions, achieving up to +0.26 SSIM and +8.3~dB PSNR improvements. Besides, Ptychi-Evolve records algorithm lineage and evolution metadata, enabling interpretable and reproducible analysis of discovered regularizers.
Read the full abstract

All authors Xiangyu Yin, Ming Du, Junjing Deng, Zhi Yang, Yimo Han, Yi Jiang

Algebraic Topology

Fully-Dualizable and Invertible $\mathcal{E}_n$-Algebras

We prove a conjecture of Brochier, Jordan, Safronov, and Snyder [BJSS21], first formulated by Lurie [Lur09b], characterizing fully-dualizable and invertible $\mathcal{E}_n$-algebras viewed as objects in the higher Morita categories $\mathbf{Mor}_n(\mathcal{V})$ [Lur09b, Sch14, Hau17b, Hau23]. In other words, we characterize those $\mathcal{E}_n$-algebras which give rise to $(n + 1)$-dimensional topological quantum field theories (TQFT), and those which give rise to invertible theories.
Read the full abstract

Combinatorics

Extremal degree-based indices of general polyomino chains via dynamic programming

In this paper, we develop a dynamic programming framework for identifying extremal general polyomino chains with respect to degree-based topological indices. As a concrete application, we resolve an open problem posed in 2015 by determining, for any given number of squares, the general polyomino chains that maximize the generalized Randić index with parameter $\alpha=-1$. We show that the extremal configurations depend explicitly on the residue class of the number of squares modulo 4. Beyond this specific result, the proposed dynamic programming approach provides a constructive and systematic methodology for tackling extremal problems in graph theory.
Read the full abstract

History and Overview of Mathematics

Blackwells Demon: Postdiction and Prediction in Random Walks

Maxwells Demon is a mythical being, first described by the physicist James Clerk Maxwell (although named Maxwells Demon by Lord Kelvin). Maxwell used it in a thought experiment to potentially violate the Second Law of Thermodynamics by exploiting inhomogeneities existing in a statistically homogeneous system. Blackwells Demon, making (as far as is known) its first appearance in this paper, illustrates a counterintuitive situation occurring in a random walk variation of the Two Envelope problem[1], that it is possible under restrictive conditions to predict with success probability > 1/2 the direction of a random walk generated by the flip of a fair coin. Like Maxwells Demon, Blackwells Demon operates by exploiting inhomogeneities that exist in a statistically homogeneous system. Maxwells Demon achieves its results by knowing when a molecule is moving rapidly and when it is not. Blackwells Demon achieves its results by knowing when a prediction strategy is successful and when it is not. At the time Maxwell proposed his Demon, it confronted a technological Everest, the ability to open and close a gate permitting the passage of a single molecule, and the ability to gauge the speed of an approaching molecule. Blackwells Demon merely has to turn a light on and off, conduct visual observations and keep simple statistical records. It should be noted that the analysis in this paper does not demonstrate the ability to predict the flip of a fair coin ab initio with success probability > 1/2, as it is necessary to embed the fair coin in an environment of some complexity in order to achieve this result.
Read the full abstract

Algebraic Geometry · Cross-listed from cs.LG

Reinforcement Learning for Power-Flow Network Analysis

The power flow equations are non-linear multivariate equations that describe the relationship between power injections and bus voltages of electric power networks. Given a network topology, we are interested in finding network parameters with many equilibrium points. This corresponds to finding instances of the power flow equations with many real solutions. Current state-of-the art algorithms in computational algebra are not capable of answering this question for networks involving more than a small number of variables. To remedy this, we design a probabilistic reward function that gives a good approximation to this root count, and a state-space that mimics the space of power flow equations. We derive the average root count for a Gaussian model, and use this as a baseline for our RL agents. The agents discover instances of the power flow equations with many more solutions than the average baseline. This demonstrates the potential of RL for power-flow network design and analysis as well as the potential for RL to contribute meaningfully to problems that involve complex non-linear algebra or geometry. \footnote{Author order alphabetic, all authors contributed equally.
Read the full abstract

Differential Geometry

Lipschitz Bounds and Uniform Convergence for Sequences of Bounded Rough Riemannian Metrics

Here we study what we call bounded rough Riemannian metrics $(M,g)$, which are positive definite, symmetric tensors on each tangent space, $T_pM$, which are bounded and measurable as functions in coordinates. This is enough structure to study the length space given by taking the infimum of the length of all piecewise smooth curves connecting points $p,q \in M$. The goal is to find the weakest conditions one can place on $g$ which can guarantee Lipschitz or uniform bounds from above and below. For each condition, an example is given showing that the condition cannot be weakened any further which also explores the geometric intuition.
Read the full abstract

All authors Brian Allen, Bernardo Falcao, Harry Pacheco, Bryan Sanchez

Geometric Topology

A homological generalized Property R conjecture is false

The generalized Property R conjecture (GPRC) predicts that if framed surgery on an $n$-component link $L$ in $S^3$ produces $\#^{n} (S^1\times S^2)$, then $L$ is handleslide equivalent to an unlink, the obvious way to construct such a surgery. Many potential counterexamples to the GPRC are known, but obstructing handleslide equivalence is a tricky proposition. In this vein, we disprove a further generalization of the GPRC. It would be reasonable to expect that if an $n$-component link in $S^3$ surgers to the connected sum of $n$ three-manifolds with the homology of $S^1 \times S^2$, then this link should be handleslide equivalent to an $n$-component split link, the obvious way to construct such a surgery. However, we prove that there are 2-component framed links in $S^3$ that surger to a connected sum of homology $S^1\times S^2$'s but that are not handleslide equivalent, or even weakly handleslide equivalent, to a split link.
Read the full abstract

Combinatorics

Graph labellings and external difference families

Digraph-defined external difference families were recently introduced as a natural generalization of several well-studied combinatorial objects motivated by cryptography (e.g. external difference families (EDFs) and circular external difference families (CEDFs)). In this paper, we develop a systematic framework for using various types of vertex-labellings for graphs and digraphs to create digraph-defined external difference families. The approach is to combine suitable vertex-labellings (generalizations of $\alpha$-valuations, namely near $\alpha$-valuations and oriented near $\alpha$-valuations) with a graph blow-up technique. Many new families are produced, including the first explicit construction for an infinite family of $2$-CEDFs, achieving all parameter sets for $(n,m,l;1)$-$2$-CEDFs with $m \equiv 0 \mod 4$ sets. Further, new results arise for graph labellings themselves (e.g. cyclotomy-based near $\alpha$-valuations for a family of trees without $\alpha$-valuations, and an $\alpha$-valuation for sun graphs).
Read the full abstract

Commutative Algebra

Algebraic Invariants of Edge Ideals Under Suspension

The central question of this paper is: how do algebraic invariants of edge ideals change under natural graph operations? We study this question through the lens of suspensions. The (full) suspension of a graph is obtained by adjoining a new vertex adjacent to every vertex of the original graph; this construction is well-understood in the literature. Motivated by the fact that regularity is preserved under full suspension while projective dimension becomes maximal, we refine the construction to selective suspensions, where the new vertex is joined only to a prescribed subset of vertices. We focus on two extremal choices: minimal vertex covers and maximal independent sets. For suspensions over minimal vertex covers of an arbitrary graph, regularity is preserved and projective dimension increases by one. Moreover, the independence polynomial changes in a controlled way, allowing us to track $\mathfrak a$-invariants under cover suspension. In contrast, the analogous uniform behavior fails in general for suspensions over maximal independent sets. We therefore analyze paths and cycles and give a complete description: projective dimension always increases by one, and regularity and the $\mathfrak a$-invariant are preserved except for a unique extremal family of paths, where both invariants increase by one.
Read the full abstract

Geometric Topology

A refined 1-cocycle for regular isotopies and the refined tangle equations

We refine the combinatorial 1-cocycle $\mathbb{L}R_{reg}$ for regular isotopies of long knots to a 1-cocycle with values in the free $\mathbb{Z}[x,x^{-1}]$-module generated by regular isotopy classes of oriented tangles with exactly one signed ordinary double point. We use it to define the refined tangle equations for couples of knot diagrams, where the coefficients are now Laurent polynomials instead of integers. A solution of the tangle equations gives quantitative information about any knot isotopy which relates two given knot diagrams. If the tangle equations have no solution, then the diagrams represent different knots.
Read the full abstract

Probability

Fluctuations for the Sherrington–Kirkpatrick spin glass model near the critical temperature

We consider the Sherrington–Kirkpatrick spin glass model with zero external field and at inverse temperature $\beta>0$. Let $F_N(\beta)$ be the corresponding log-partition function. Under the assumption that $c_N:=N^{1/3}(1-\beta_N^2)$ is bounded away from $0$, we prove that Var$(F_N(\beta_N)) = - \frac{1}{2} \log (1-\beta_N^2) -{\beta_N^2}/{2} + O( c_N^{-3/2}).$ As a consequence, we obtain Var$(F_N(1-c N^{-1/3})) = \frac16\log N + O(1)$ for any fixed constant $c\in(0,\infty)$. We also prove a Gaussian central limit theorem for the centered and scaled $F_N(\beta_N)$.
Read the full abstract

Rings and Algebras

On nonmatrix varieties of associative rings

We study nonmatrix varieties of $\mathbf{k}$-algebras, where $\mathbf{k}$ is a unital commutative ring. Our results extend to this generality known results for the case in which $\mathbf{k}$ is an infinite field. Also, we generalize these results to varieties of $\mathbf{k}$-algebras not containing the algebra of $n\times n$ matrices.
Read the full abstract

Information Theory · Cross-listed from eess.SP

A Survey on Stacked Intelligent Metasurfaces: Fundamentals, Recent Advances, and Challenges

Stacked intelligent metasurfaces (SIMs) cascade several closely spaced programmable metasurface layers so that an electromagnetic wave undergoes repeated modulation-propagation interactions before leaving the stack. The stack thereby implements, at the speed of propagation and with few radio-frequency chains, high-dimensional analog transformations that would otherwise require digital precoding, combining, spatial filtering, focusing, or inference. Since 2023, SIM research has produced more than two hundred articles, a dedicated magazine special issue, and two survey-level overviews organized around applications and wave-domain computing. This article complements them with a model-to-hardware perspective. First, we present the electromagnetic modeling hierarchy of SIMs, making explicit the assumptions under which the models coincide, their validity regimes, and their computational complexity, with a common notation for hardware non-idealities and power consumption. Second, we present a methodology-first taxonomy (SIM-aware channel acquisition, model-based optimization, statistical-CSI design, learning-based control, and hardware-aware robust design), compare the families in terms of CSI requirements, complexity, convergence, and hardware constraints, and reconcile conflicting findings in the literature. Third, we synthesize applications by the physical role of the SIM. Fourth, we quantitatively compare experimentally demonstrated multilayer prototypes, separate validated from conceptual architectures, and discuss thermal, tolerance, and calibration issues. Fifth, we provide a tutorial-style performance and trade-off analysis anchored in reported results. Finally, we map evidence-based challenges to research directions with measurable targets. Lessons learned are summarized at the end of every section.
Read the full abstract

All authors Chandan Kumar Sheemar, Sourabh Solanki, Wali Ullah Khan, George C. Alexandropoulos, Symeon Chatzinotas

Dynamical Systems · Cross-listed from q-bio.PE

The Impact of Neglecting Vaccine Unwillingness in Epidemiology Models

With significant population fractions in many societies who refuse vaccines, it is important to reconsider how vaccination is incorporated into compartmental epidemiology models. It is still most common to apply the vaccination rate to the entire class of susceptibles, rather than to use the more realistic assumption that the vaccination rate function should depend only on the population of susceptibles who are willing and able to receive a vaccination. This study uses a simple generic disease model to address two questions: (1) How much error is introduced in key model outcomes by neglecting vaccine unwillingness?, and (2) Can the error be reduced by incorporating vaccine unwillingness into the vaccination rate constant rather than the rate diagram? The answers depend greatly on the time scale of interest. For the endemic time scale, where longterm behavior is studied with equilibrium point analysis, the error in neglecting unwillingess is large and cannot be improved upon by decreasing the vaccination rate constant. For the epidemic time scale, where the first big epidemic wave is studied with numerical simulations, the error can still be significant, particularly for diseases that are relatively less infectious and vaccination programs that are relatively slow.
Read the full abstract

Optimization and Control

Mean-field games with unbounded controls: a weak formulation approach to global solutions

We establish an existence of equilibrium result for a class of non-Markovian mean-field games with unbounded control space in weak formulation. Our result is based on new existence and stability results for quadratic-growth generalized McKean-Vlasov BSDEs. Unlike earlier approaches, our approach does not require boundedness assumptions on the model parameters or time horizons and allows for running costs that are quadratic in the control variable.
Read the full abstract

Optimization and Control

Decision-dependent distributionally robust standard quadratic optimization with Wasserstein ambiguity

The standard quadratic optimization problem (StQP) consists of minimizing a quadratic form over the standard simplex. Without assuming convexity or concavity of the quadratic form, the StQP is NP-hard. This problem has many interesting applications ranging from portfolio optimization to machine learning. Sometimes, the data matrix is uncertain but some information about its distribution can be inferred, e.g. a distance to a reference distribution (typically, the empirical distribution after sampling). In distributionally robust optimization, the goal is to hedge against the worst case of all possible distributions in an ambiguity set, defined by above mentioned distance. In this paper we will focus on distributionally robust StQPs under Wasserstein distance, and show equivalence to an accordingly modified deterministic instance of an StQP. This blends well into recent findings for other approaches of StQPs under uncertainty. We will also address out-of-sample performance guarantees. Carefully designed experiments shall complement and illustrate the approach.
Read the full abstract

All authors Immanuel M. Bomze, Daniel de Vicente, Abdel Lisser, Heng Zhang

Number Theory

Joint Linnik problems

We prove a conjecture of Michel–Venkatesh on joinings of distinct Linnik problems, in the setting of simultaneous quaternionic embeddings of imaginary quadratic fields having sufficiently many small split primes. This splitting condition is known to hold for all but $O((\log\log X)^{1 + o(1)})$ discriminants up to $X$. We also treat a non-equivariant form of this conjecture proposed by Aka–Einsiedler–Shapira, which in particular applies to the classical Gauss construction joining Linnik points on the sphere with CM points on the modular surface.
Read the full abstract

Classical Analysis and ODEs

On the approximation of Weierstrass function via superoscillations

The Weierstrass function is a classic example of a continuous nowhere differentiable function, defined as a sum of high-frequency complex exponentials. In this paper, we follow a suggestion of M.V. Berry and study the convergence properties of Berry's superoscillating approximation to the truncated Weierstrass function. We provide sharp, explicit error estimates for this approximation and we analyze the subtle convergence properties of the associated double limits.
Read the full abstract

Optimization and Control · Cross-listed from cs.LG

A Novel Hybrid Heuristic-Reinforcement Learning Optimization Approach for a Class of Railcar Shunting Problems

Railcar shunting is a core planning task in freight railyards, where yard planners need to disassemble and reassemble groups of railcars to form outbound trains. Classification tracks with access from one side only can be considered as stack structures, where railcars are added and removed from only one end, leading to a last-in-first-out (LIFO) retrieval order. In contrast, two-sided tracks function like queue structures, allowing railcars to be added from one end and removed from the opposite end, following a first-in-first-out (FIFO) order. We consider a problem requiring assembly of multiple outbound trains using two locomotives in a railyard with two-sided classification track access. To address this combinatorially challenging problem class, we decompose the problem into two subproblems, each with one-sided classification track access and a locomotive on each side. We present a novel Hybrid Heuristic-Reinforcement Learning (HHRL) framework that integrates railway-specific heuristic solution approaches with a reinforcement learning method, specifically Q-learning. The proposed framework leverages methods to decrease the state-action space and guide exploration during reinforcement learning. The results of a series of numerical experiments demonstrate the efficiency and quality of the HHRL algorithm in both one-sided access, single-locomotive problems and two-sided access, two-locomotive problems.
Read the full abstract

Probability · Cross-listed from cs.LG

Autocorrelation effects in a stochastic-process model for solving two-armed bandit problems

Decision makers exploiting photonic chaotic dynamics obtained by semiconductor lasers provide an ultrafast approach to solving multi-armed bandit problems by using a temporal optical signal as the driving source for sequential decisions. In such systems, the sampling interval of the chaotic waveform shapes the temporal correlation of the resulting time series, and experiments have reported that decision accuracy depends strongly on this autocorrelation property. However, it remains unclear whether the benefit of autocorrelation can be explained by a minimal mathematical model. Here, we analyze a stochastic-process model for solving the two-armed bandit problem based on time series, where the threshold and a two-valued Markov signal evolve jointly. Numerical results reveal an environment-dependent structure: negative (positive) autocorrelation is optimal in reward-rich (reward-poor) environments. These findings show that negative autocorrelation of the time series is advantageous when the sum of the winning probabilities is more than one, whereas positive autocorrelation is useful when the sum of the winning probabilities is less than one. Moreover, the performance is independent of autocorrelation if the sum of the winning probabilities equals one, which is mathematically clarified. This study paves the way for solving the two-armed bandit problems for reinforcement learning applications in wireless communications and robotics.
Read the full abstract

All authors Tomoki Yamagami, Mikio Hasegawa, Takatomo Mihana, Ryoichi Horisaki, Atsushi Uchida

Analysis of PDEs

Sobolev regularity of the symmetric gradient of solutions to a class of $\phi$-Laplacian systems

The paper deals with the second order regularity properties of the weak solutions $u\in W^{1,\phi}(\Omega, \real^n)$ } of systems of the form \begin{equation*}\label{equareg} -\dive A(x,\E u)=f, \end{equation*} in a bounded domain $\Omega\subset\R^n$, $n>2$, where the operator $ A(x,P)$ is Lipschitz continuous with respect to the $x$-variable and satisfies growth conditions with respect to the second variable expressed through a Young function $\Phi$. We prove the Sobolev regularity of a function of the symmetric gradient $\E u$ that takes into account the nonlinear growth of the operator $A(x,P)$, {assuming that the force term $f$ belongs to a suitable Orlicz-Sobolev space. {The main result is achieved through some uniform higher differentiability estimates for solutions to a class of approximating problems, constructed adding singular higher order perturbations to the system.
Read the full abstract

Algebraic Geometry

Reductification of parahoric group schemes

Parahoric group schemes are certain possibly non-reductive, smooth, affine integral models of reductive group schemes defined over a henselian discretely valued field $K$ whose residue field is perfect. We show that any such group scheme $\mathscr{P}$ becomes reductive, in a particular regard, after a (possibly wildly ramified) finite Galois extension $L/K$. More precisely, we prove that there exists a reductive integral model $\mathscr{G}$ of the base change $\mathscr{P}_L$ such that $\mathscr{P}$ can be recovered as the smoothening of the subgroup of Galois invariants of the Weil restriction of $\mathscr{G}$. Our work extends results of Balaji–Seshadri and Pappas–Rapoport from the tamely ramified and simply-connected semisimple setting. As an application, we establish a parahoric analogue of the Grothendieck–Serre conjecture in sufficiently good residue characteristics. Specifically, we confirm that generically trivial parahoric torsors are trivial whenever the generic reductive group is simply-connected. The proof proceeds by reducing the problem to a statement about a stacky reductive group over a stacky discrete valuation ring.
Read the full abstract

Number Theory

Waring-Goldbach problems for one square and higher powers

We prove that every sufficiently large odd integer can be expressed as a sum of one square and fourteen fifth powers, all of primes. In addition, we establish that every sufficiently large even integer can be written as a sum of one square, one biquadrate, and twelve fifth powers of primes.
Read the full abstract

Number Theory

The fourth known primitive solution to $a^5 + b^5 + c^5 + d^5 = e^5$

We report the fourth known primitive solution to the Diophantine equation $a^5 + b^5 + c^5 + d^5 = e^5$, extending the small set of solutions identified in 1966, 1996, and 2004. The new solution has target value $e = 1{,}956{,}878$, more than an order of magnitude larger than those of all previously known primitive solutions. It was obtained by a large-scale computational search using an optimized meet-in-the-middle algorithm. To make this search feasible under fixed memory, we introduce a partitioning strategy that trades additional preprocessing for reduced memory usage, and show that this approach remains practical for search ranges well beyond those explored here. Finally, we compare structural properties of known fourth- and fifth-power counterexamples to Euler's sum-of-powers conjecture, observing that most known fifth-power solutions have their largest summand close to the target value.
Read the full abstract

Logic

Capturing dual team properties with inclusion atoms

We introduce propositional team-based logics expressively complete for (quasi) downward and (quasi) upward closed properties in a syntactically dual way, by using variants of the inclusion atom. In particular, the variants of the primitive inclusion atoms used in the (quasi) upward closed setting have equivalent formulas using variants of the might modality. The duality is visible in the logics' normal forms, mirroring the duality between the (quasi) upward and downward closed settings, where the quasi variants take special care of the empty and full team. Furthermore, we defined sound and complete natural deduction systems for each logic.
Read the full abstract

Optimization and Control · Cross-listed from cs.LG

Cheap Thrills: Effective Amortized Optimization Using Inexpensive Labels

To scale optimization and simulation, prior work has explored training machine-learning surrogates that map problem parameters to solutions inexpensively at inference time. Unfortunately, commonly used approaches, including supervised and self-supervised learning with either soft or hard feasibility enforcement, face inherent challenges such as reliance on expensive high-quality labels or difficult optimization landscapes. To address their trade-offs, we propose a novel framework that collects "cheap" imperfect labels, performs supervised model pretraining with a merit loss-based termination scheme, and finally refines the model through self-supervised learning to improve final performance. Empirical validation across challenging domains – including nonconvex constrained optimization, power-grid operation, and stiff dynamical systems – shows that this three-stage strategy yields faster convergence; improved accuracy, feasibility, and optimality; and up to 59x reductions in total offline computational cost. We further analyze why and when our framework improves surrogate model training, finding that (i) merit loss is an informative signal and (ii) only small numbers of cheap, inexact labels are needed to place the model in a favorable regime for self-supervised learning.
Read the full abstract

All authors Khai Nguyen, Petros Ellinas, Anvita Bhagavathula, Priya L. Donti