Peer Revenue Past citations are no guarantee of future returns

Papers

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

All of arXiv

Saturday 12 September

Machine Learning

Does Reasoning Improve Psychological Depth in Large Language Models? It Depends on Who's Judging

LLM-as-a-Judge evaluators are increasingly used to score open-ended generation, yet a judge's correlation with human ratings on its development set may not guarantee valid measurement when outputs are closely matched and human preferences are subjective. We study this failure mode through psychological depth in short stories. Seven human readers and an LLM-judge ensemble selected on the original scalar Psychological Depth Scale dataset ($\rho = 0.646$) evaluated 60 blinded, prompt-matched story pairs from GPT-5 vs.\ GPT-4o and DeepSeek-R1 vs.\ DeepSeek-V3. Human preferences showed no universal reasoning advantage: GPT-5 was modestly preferred over GPT-4o (60.0–62.9%), whereas DeepSeek-R1 trailed V3 (42.9%), and inter-reader agreement was near chance (Krippendorff's $\alpha = 0.070$), with within-reader consistency and recurring weighting patterns suggesting structured heterogeneity rather than random responding. The judge, by contrast, favored reasoning outputs in 89.0% of dimension-level comparisons and 59 of 60 pairs on aggregate PDS, uniformly across all five evaluator configurations, and its scores were associated with surface features such as sentence length and lexical diversity. These results suggest that development-set performance is insufficient evidence for deployment validity on a shifted distribution, and that point-estimate judges can obscure the heterogeneity in subjective human evaluation.
Read the full abstract

All authors Ruichen Zheng, Yihe Wang, Fabrice Y Harel-Canada, Sara Khosravi, Zeynep Senahan Yildiz, Amit Sahai, Nanyun Peng

Combinatorics

Eccentric Connectivity Index of Cartesian product and Strong product of Strongly Connected Digraphs

Let $G=(V,E)$ be a graph. The eccentric connectivity index of $G$ is defined as $$\xi^C(G)=\sum_{u\in V(G)}d_uecc(u)$$ where $d_u$ and $ecc(u)$ are the degree and eccentricity of $u$, respectively. For a strongly connected digraph $D=(V,A)$, the eccentric connectivity index is defined as $$\xi^C(D)=\frac{1}{2}\sum_{u\in V(D)}(d_u^++d_u^-)mecc(u)$$ where $d_u^+$ and $d_u^-$ are the out-degree and in-degree of $u$, respectively, and $mecc(u)$ is its m-eccentricity with respect to the maximum distance $md(u,v)=\max\{\vec d(u,v),\vec d(v,u)\}$. In this article, give formulas and bounds for the eccentric connectivity index of Cartesian and strong products of strongly connected digraphs and discuss the corresponding equality cases. Also, an attempt is made to study the self-centeredness of these products and establish conditions under which the Cartesian and strong products are self-centered.The results are extended to products of several digraphs.
Read the full abstract

Artificial Intelligence

Homeostatic Continual Learning

In this paper, I formulate a Continual Learning problem and propose a method named "Homeostatic Continual Learning" that enables an AI agent to learn continuously in a changing environment without catastrophic forgetting. The core of the method is to find outliers in the environment data when the agent experiences an outlier in its output. Through this method, the agent gradually completes its model and policy and performs well in more and more contexts. I also suggest that we may use the method to build a world model where the agent factorizes the objects in the world into features, abstract objects into comparable instances of concepts and map concepts to intents through features. I discuss the works needed to render the method practical, the connections to many fields in Artificial Intelligence and the broader implications of the method.
Read the full abstract

Machine Learning

Training Specialist Models without Reasoning Trajectories for Domain Expert Distillation

Specialist distillation effectively transfers domain expertise to student models via teacher-generated reasoning trajectories. However, when these specialists are trained solely on question–answer pairs without explicit reasoning supervision, what governs the trajectories they generate? In this work, we show that specialist optimization implicitly selects from this latent trajectory space. To isolate and observe this latent distribution, we leverage student distillation not as a downstream goal, but as an agnostic probe—since students inherit no parameterization or optimization constraints from the specialist, inheriting only the sampled trajectories themselves. Through this probe, our empirical analysis unveils a tight governing relationship: across 27 specialist–student pairings, their specialization–generalization profiles correlate exceptionally strongly. Crucially, explicitly controlling the specialist's distributional drift systematically shifts both the teacher and its distilled student along a controllable trade-off between domain precision and general-capability retention. Across chemistry, physics, and multilingual settings, distilled students systematically reflect these specialist-induced profiles, even across divergent model families. Our findings establish a new view of specialist training: when gold reasoning is absent, tuning choices directly control the latent supervision passed to downstream models.
Read the full abstract

All authors Yilei Tu, Zihao Li, Shaoxiong Ji, Jörg Tiedemann, Fei Yuan

Exactly Solvable and Integrable Systems

Finite dimensional reductions of integrable differential-difference equations

Integrable partial differential equations, such as the Korteweg–De Vries (KdV) equation, admit infinite hierarchies of commuting higher symmetries. Their symmetry reductions give rise to integrable finite-dimensional dynamical systems solvable in terms of Abelian functions. This underlies the finite-gap integration method for the KdV equation, introduced by S.P. Novikov and subsequently extended to many other integrable systems. In this paper, I introduce the concept of reduction ideals and a new class of reductions for integrable differential–difference equations, leading to integrable finite-dimensional systems in both commutative and noncommutative settings. This class includes periodic reductions, symmetry reductions, and a broad family of other reductions defined in terms of the associated Lax operator. The reduction constraints define integrable maps that enable solutions of the reduced systems to be extended to solutions of the corresponding differential-difference equations. The construction is illustrated using the Volterra hierarchy.
Read the full abstract

Computation and Language

HyperProve: Answer-Guided Hypergraph Expansion for Multi-Hop Question Answering

Multi-hop question answering often fails when retrieval treats evidence as isolated matches to the original question, since the facts needed to answer a complex question are usually connected through intermediate entities, relations, and constraints. We propose HyperProve, a retrieval-augmented QA framework that addresses this challenge by coupling question decomposition with answer-conditioned expansion over a hypergraph of atomic facts. HyperProve does not use atomic facts, hypergraphs, or iterative retrieval in isolation; instead, it carries intermediate answers and supporting hyperedges as retrieval state, then uses that state to bias the next local hypergraph expansion. This design enables HyperProve to construct coherent evidence chains for final answer generation while making the retrieval process stateful and fact-centered. Across multi-hop QA benchmarks, HyperProve achieves the best overall performance in our evaluation, outperforming the strongest baselines by an average relative improvement of 6.2% in answer accuracy and 4.9% in F1.
Read the full abstract

All authors An Nguyen Phu, Dung Nguyen Quang, Luu Hieu An, Linh Ngo Van, Trung Le, Thien Huu Nguyen

Rings and Algebras

Dedekind Poisson Algebras over Arbitrary Fields

We study Poisson algebras in which every Poisson subalgebra is a Poisson ideal, called Dedekind Poisson algebras, over arbitrary fields and without any finite-dimensionality assumption. In characteristic different from 2, combining the two operations by $x*y=xy+[x,y]$ connects the problem with Outcalt's classical classification of power-associative H-algebras and yields the same structural type. We obtain a complete classification by a direct two-operation argument valid in every characteristic. This approach also shows that characteristic 2 retains additional Poisson information which cannot be recovered from the single product $*$.
Read the full abstract

Quantum Algebra

Cyclotomic expansions of colored $SU(n)$ invariants of two-strand torus knots and Bailey transforms

Habiro's cyclotomic expansion of the colored Jones polynomial has a higher rank analogue conjectured by Chen–Liu–Zhu for colored $SU(n)$ invariants. We prove this conjecture for torus knots $T(2,2p+1)$ and give a Bailey theoretic realization of the cyclotomic coefficients. For $n\geq2$ and $K_p=T(2,2p+1)$, the colored $SU(n)$ invariants admit an expansion $$ J_N^{SU(n)}(K_p;q) = \sum_{m=0}^{N} \left(\prod_{j=0}^{m-1}\{N-j\}\{N+n+j\}\right) H_m^{(n,p)}(q), $$ where $H_m^{(n,p)}(q)\in\mathbb Z[q^{\pm1}]$ is independent of the color $N$. The proof identifies the cyclotomic basis with a Newton basis and rewrites the resulting Newton coefficients as an ordinary Bailey transform. The Lin–Zheng formula for $T(2,2p+1)$ then gives a well-poised Bailey kernel. A terminating very-well-poised ${}_6\phi_5$ summation diagonalizes the kernel and reduces the integrality problem to an ordinary Bailey transition. We prove a uniform integrality theorem for these transitions in a formal integral $q$-difference operator algebra. As a direct corollary, the expansion yields the corresponding congruence relations and proves part(i) of the Chen–Liu–Zhu $SU(n)$ volume conjecture for $T(2,2p+1)$.
Read the full abstract

Optimization and Control

Last-Iterate Performance of Gradient Descent and Relaxed Proximal Point via $s$-Composability

The notion of s-composability was introduced to compose optimized stepsize schedules while preserving sharp guarantees. We show that the same joint potential has a second use: it can serve directly as a certificate of sharp last-iterate performance. We develop this viewpoint for constant and silver schedules. For the constant schedule, we prove the s-composability statement posed as an open question by Grimmer et al. (2025), at every finite horizon, through an explicit nonnegative smooth-convex interpolation certificate. The result identifies the unique constant stepsize minimizing the worst-case final gradient norm of smooth convex gradient descent under an initial-distance bound, thereby proving the optimal stepsize, value, and uniqueness predicted by Conjecture 3 of Taylor et al. (2017) without resolving its full worst-case curve. Through the Moreau envelope, the same constant is also the unique constant relaxation minimizing the worst-case final residual of relaxed proximal point. More generally, for every positive s-composable schedule, we determine the exact worst-case final proximal objective-gap constant and give a matching one-dimensional example. Applying this result to the already s-composable original silver schedule yields its exact last-iterate objective-gap constant, closing a question left open by Wang et al. (2025).
Read the full abstract

Combinatorics

Coincidences and Growth of Boxed Mesh Patterns

A boxed mesh pattern is a mesh pattern whose selected entries lie in an empty axis-parallel rectangle. We classify coincidences of boxed patterns with classical and vincular patterns, exhibit a genuinely bivincular coincidence, and prove that no boxed–bivincular coincidence occurs for patterns of length at least five. Together with known results, this shows that every boxed pattern of length at least five has factorial growth and hence fails the Stanley–Wilf property. At length four, one exceptional orbit is enumerated by the semi-Baxter numbers, while the remaining exceptional orbit, $\{2143,3412\}$, is unresolved; we conjecture that it has factorial growth. For Box(123), we derive an exact maximum-insertion identity and prove the subfactorial upper bound $2^{5n}n^{\beta n}$, where $\beta=\log_2(2\cos(\pi/7))<0.85$. A closed enumeration remains open. We also prove a general first-moment formula for boxed mesh patterns that depends only on the length of the underlying pattern; in particular, the expected number of Box(123) occurrences in a uniformly random permutation of length $n$ is asymptotic to $n\log n$. For Box(12), we identify the occurrence statistic with the up-degree in the strong Bruhat order, obtaining its maximum, its mean, and an exact insertion identity for the distribution polynomials. We conjecture that the coefficients of these polynomials are unimodal.
Read the full abstract

Analysis of PDEs

Rigidity and stability for the Bogovskii constant

We prove that balls are the unique minimisers of the Bogovskii constant among bounded connected Lipschitz domains in $\mathbb{R}^n$, $n\geq 2$, resolving Open Problem 2 of Gazzola, Grunau and Sperone (arXiv:2608.25861).

Quantum Gases

Surrogate-assisted three-dimensional Gross-Pitaevskii characterization of Bose-Einstein condensate transport on an atom chip

We develop a surrogate-assisted workflow for millimeter-scale transport of a $^{87}$Rb Bose–Einstein condensate (BEC) on a multi-wire atom chip. Gradient-based inverse optimization generates 42 current schedules spanning seven regularization strengths and six transport durations. A low-cost Gaussian phase-space model ranks these schedules by endpoint overlap at three atom numbers. A Thomas–Fermi scaling (Ermakov) model estimates cloud dimensions for case-specific moving-grid construction. We then evaluate selected schedules in detail with the three-dimensional Gross–Pitaevskii equation (GPE). At $N=10^3$ and $T=0.5$ s, the surrogate predicts a larger axial center-of-mass (COM) excursion than at longer durations. The GPE confirms this prediction and reveals sloshing that persists through delivery, with an endpoint fidelity $F_{3D}=0.9244$. For all five longer transports, the surrogate predicts near-unit endpoint overlap for the top-ranked schedules, and GPE propagation using these schedules gives $F_{3D}>0.9997$. At $N=10^4$ and $T=2.0$ s, the surrogate-selected schedule similarly gives $F_{3D}=0.999932$. These results demonstrate that the surrogate estimates provide a useful low-cost basis for candidate screening, while GPE propagation resolves the transport dynamics and the delivered condensate state.
Read the full abstract

Robotics

Learning In-Hand Object Reaching to General 6D Poses

In-hand manipulation allows multi-fingered dexterous hands to reconfigure grasped objects without releasing and regrasping them. This improves manipulation efficiency by reducing repeated grasp acquisition and large arm motions. However, most learning-based methods focus on reorientation, continuous rotation, or translation, whereas many tasks require joint control of object position and orientation. We formulate this capability as in-hand 6D object pose reaching: starting from an existing grasp, coordinated finger motions move the object to a palm-relative target pose. We present POISE (Palm-relative Object reaching In SE(3)), a sim-to-real reinforcement learning framework for this task. POISE combines diverse stable-grasp initialization, goal- and geometry-conditioned control, an adaptive 6D goal curriculum, and a compact reward scheme for pose reaching and grasp preservation. In simulation, diverse initialization raises held-out-grasp success from 40.1% to 51.5% and post-drop recovery from 33.8% to 72.9%; the curriculum raises full-range success from 6.2% to 59.5%. On hardware, the grasp-maintenance reward improves three-target sequence success from 20% to 80%. In real-world experiments, POISE reaches successive 6D targets without manual reset across multiple object geometries and wrist orientations, and recovers from external disturbances. To support further research in dexterous manipulation, we will release our code at https://junxiaolin.github.io/poise-website/.
Read the full abstract

All authors Junxiao Lin, Tianyue Wu, Jie Yin, Jia Pan, Kaifeng Zhang, Weiming Zhi

Artificial Intelligence

Positioning manuscripts in the scientific landscape with agentic AI

Publishing a research manuscript is a routine yet demanding part of scientific life: time-consuming, stressful, and often uncertain in outcome. Recent advances in large language model (LLM)-based agentic AI have shown promise across a range of scientific tasks, and here we ask whether agentic AI can help researchers navigate the publication process itself by reliably inferring a manuscript's eventual publication venue from its content and literature context. We introduce PASS (Publication-oriented Agentic Scientific System), an agentic system that understands manuscripts within their domain-specific literature context and predicts top-matched publication venues. PASS positions each manuscript within its surrounding literature landscape by reconstructing its local scientific neighborhood, tracing its topic trajectory, and reasoning over field-specific journal spaces. Evaluated on a leakage-audited benchmark of over 2,000 preprints across 16 biomedical fields, PASS achieved Top-1 accuracy of 50.3% and Top-5 accuracy of 86.1%, outperforming state-of-the-art LLM baselines and established journal-selection tools. PASS-produced quality scores, such as impact potential and novelty, aligned with independent measures of publication outcome. We also found that the designed literature retrieval module is the strongest performance contributor, particularly for positioning manuscripts relative to nearby work, and that PASS maintained near-full performance from the abstract alone, whereas LLM baselines required the full manuscript text. An independent human evaluation found strong researcher agreement with PASS's manuscript understanding and recommendation rationale. PASS has been released as a public platform (https://ratemypaper.ai/) for broad researcher access.
Read the full abstract

All authors Jiawen Chen, Zichen Zhang, Bingxuan Li, Quan Sun, Yiyan Zhang, Edric Tam, Jinjie Lin, Didong Li, Yun Li, Bingxin Zhao

Nuclear Theory

Nuclear landscape based on point-coupling density functional with localized exchange terms

Nuclear landscape is initially explored in the framework of relativistic Hartree-Bogoliubov theory under the spherical approximation, adopting the newly developed PCF-PK1 density functional. The functional effectively incorporates exchange terms via the Fierz transformation and explicitly includes the tensor coupling. We analyze the limits of the nuclear landscape and nuclear ground state properties including binding energies, charge radii, {\alpha}-decay energies, compared with other functionals and experiment. In the present calculations, 7210 nuclei are predicted to be bound with the root-mean-square deviation of binding energies 7.170 MeV. The removal of spurious shell closure at Z = 58 and 92 is discussed by shell gaps and single-particle spectra. For superheavy nuclei, potential magic numbers beyond 208Pb are studied. The inclusion of the tensor coupling in PCF-PK1 helps restore the pseudospin symmetry, leading to a less pronounced shell closure at Z = 120.
Read the full abstract

Optimization and Control

On the Equivalence of Stochastic Control and Path Space Formulations for Schrödinger Bridges over Compact Connected Lie Groups

We establish the equivalence between the stochastic optimal control and path space formulations of the Schrödinger bridge problem (SBP) for the kinematic equation on a compact connected Lie group. Using the geometric concepts of horizontal lift and stochastic anti-development, we derive a Girsanov-type change-of-measure result, and show that the expected control energy equals the relative entropy of the controlled path law with respect to the reference Wiener measure. Thus, the SBP is equivalently a path space relative entropy minimization problem subject to prescribed endpoint marginals.Our result has three useful implications. From an analytic viewpoint, the shown equivalence helps prove the existence and uniqueness of the SB. From a probabilistic viewpoint, it helps interpret the SB as the most probable deviation of the uncontrolled stochastic dynamics consistent with the endpoint constraints. From a computational viewpoint, it allows using static Sinkhorn recursions to directly solve the relative entropy minimization problem and compute the optimal path measure. We illustrate the equivalence numerically on the torus $\mathbb{T}^2$. The code is publicly available at: https://github.com/gradslab/LargeDeviationSBP
Read the full abstract

All authors Hamza Mahmood, Georgiy A. Bondar, Abhishek Halder, Adeel Akhtar

Combinatorics

The crossing number and the unit-distance crossing number of the Hamming graphs H(d,3)=K_3^{box d}, and their realizations over many coordinate fields

The Hamming graph H(d,3)=K_3^{box d} (n=3^d vertices) is the graph of single-symbol errors of ternary codes. Can it be drawn in the plane with every edge one unit long and no two non-adjacent vertices a unit apart? It can, for every d, joining two problems into one. Arithmetic: over which coordinate fields is it realizable? As a Minkowski sum of unit triangles (planar for the whole family only in base 3), each H(d,3) has not one faithful realization but a (d-1)-parameter continuum, whose field climbs the constructibility hierarchy (ruler-and-compass, origami, and beyond); yet almost every realization is transcendental, in no number field, matching the exists-R hardness of unit-distance recognition. Combinatorial: beside the ordinary crossing number cr we introduce the unit-distance crossing number udcr(G), the fewest crossings over faithful unit drawings (every edge a unit segment); it is the rectilinear crossing number under unit length, so cr <= rcr <= udcr. We prove cr(H(d,3))=Theta(n^2) (one-page 7/6 n^2 above; spectral bisection Omega(n^2) below) and cr <= 2/3 n^2 via a two-page book layout; and udcr(H(3,3))=231 exactly, with udcr(H(d,3))=O(n^2 log n). A structure theorem makes every faithful planar realization a Minkowski sum, so the realization variety is an explicit (d-1)-parameter flex containing the udcr optimum; the matching udcr=Omega(n^2 log n) is the central open problem. The invariants differ: for cr the crossing-minimal drawings are structurally opaque, whereas the udcr optimum is itself a faithful Minkowski realization. All claims are verified in exact arithmetic with reproducible code and figures.
Read the full abstract

2609.13757 · 12 Sep · v2: 49 pages. Introduces the unit-distance crossing number udcr as a defined invariant (now stated in the abstract); adds a two-page book bound cr(H(d,3)) <= 2/3 n^2, verified free rectilinear drawings cr(H(3,3)) <= 99 and cr(H(4,3)) <= 1817, spectral references (Fiedler; Brouwer-Haemers; Mohar), and a Conclusions section separating theorems from conjectures. Reproducible code and figures

Computer Vision

Development of Low-Cost Real-Time Driver Drowsiness Detection System using Eye Centre Tracking and Dynamic Thresholding

One in every five vehicle accidents on the road today is caused simply due to driver fatigue. Fatigue or otherwise drowsiness, significantly reduces the concentration and vigilance of the driver thereby increasing the risk of inherent human error leading to injuries and fatalities. Hence, our primary motive being - to reduce road accidents using a non-intrusive image processing based alert system. In this regard, we have built a system that detects driver drowsiness by real time tracking and monitoring the pattern of the driver's eyes. The stand alone system consists of 3 interconnected components - a processor, a camera and an alarm. After initial facial detection, the eyes are located, extracted and continuously monitored to check whether they are open or closed on the basis of a pixel-by-pixel method. When the eyes are seen to be closed for a certain amount of time, drowsiness is said to be detected and an alarm is issued accordingly to alert the driver and hence, prevent a casualty.
Read the full abstract

Optimization and Control

Row-Polar LP-Newton for Linear Programming with Corral Repair

LP-Newton solves a linear program through a sequence of nearest-point problems. Given an interior feasible point for an inequality-form LP, we construct the compact hull formed by the origin and its normalized constraint rows. Polar LP-Newton (P-LPN) follows the objective ray to the boundary of this row-polar hull, where it recovers a primal-dual optimum or a recession direction proving unboundedness. For rational row-polar data, we prove an outer-iteration bound quadratic in dimension and linear in binary input length, excluding inner Wolfe work. Restarting Wolfe at every outer iteration discards its terminal corral even though the hull is unchanged and the next target lies on the same ray. Corral-repair P-LPN (CR-P-LPN) instead repairs that corral and uses it to start the next projection. Verification over the full hull preserves P-LPN's projections, outer targets, and LP conclusion in exact arithmetic. Experiments in Julia and MATLAB show where repair helps. After common initialization, median P-LPN times across 180 single-LP tests are 1.62 times the corresponding CR-P-LPN times in Julia and 1.55 times in MATLAB. On three problems with 5,000 to 100,000 rows, the corresponding geometric-mean factors are 2.03 and 1.46. A separate end-to-end study includes initialization on 19 application-derived LPs. CR-P-LPN has lower median total time than the faster cold HiGHS mode, either dual simplex or the interior-point method with crossover, on 18 of the 19 workloads in each language. The geometric means of the ratios of HiGHS time to CR-P-LPN time are 4.93 and 3.06. Component comparisons identify terminal-corral reuse as the useful component, whereas candidate-first ordering alone shows no consistent gain. Repair is not always beneficial: it is slower on Klee-Minty tests, and an author-implemented simplex is faster on the retained Netlib models.
Read the full abstract

Optimization and Control

Preconditioned Three-Term Conjugate Gradient Framework for Nonconvex Finite Minimax Problems

This paper develops a hyperbolic-majorization preconditioned three-term nonlinear conjugate-gradient framework for nonconvex finite minimax optimization. An analytic symmetric positive definite metric is derived from a global quadratic majorization of the hyperbolic smoothing model and is used simultaneously as a curvature absorber, a preconditioner, and the line-search energy metric. With the displacement \(s_{k-1}\) fixed and a variable curvature response \(b_k\), an enhanced three-term direction is introduced together with an adaptive parameter \(\mu_k^\star\) that is maximal under the requirement that the baseline worst-case metric-energy constant be preserved. The resulting framework yields a Dai–Liao-type conjugacy relation, enhanced sufficient descent, a smoothing-parameter-uniform Armijo lower bound, fixed-smoothing global first-order convergence and complexity, and Clarke-stationary accumulation points under continuation.
Read the full abstract

Analysis of PDEs

A Nonlocal Perfusion-Gated Angiogenesis System: Global Weak Solutions, Fast-Signal Limits, and Invasion Fronts

Continuum angiogenesis systems often use local endothelial or vessel density as an oxygen-delivery proxy without distinguishing structurally formed vessels from pressure-supported vascular function. We formulate a two-dimensional model in which lumenized density determines a normalized nonlocal conductivity, a globally solved pressure field, and a flow-functional density that gates oxygen delivery and vessel regression. For fixed positive regularization parameters and signal relaxation times, we prove Lipschitz stability of the pressure–perfusion map and global bounded weak solvability; strong vessel compactness follows from a nonlocal ordinary differential equation stability estimate rather than spatial smoothing. As the oxygen and vascular endothelial growth factor timescales vanish simultaneously, weak solutions converge along a subsequence to a parabolic–elliptic–ordinary differential equation system, with both fast fields converging strongly in \(L^2(0,T;H^1)\). A locally frozen one-dimensional reduction admits monotone fronts precisely at or above the threshold \(B+2\sqrt{DR}\) for \(B\geq-\sqrt{DR}\) and \(-DR/B\) otherwise. Finite-volume experiments verify the analytical mechanisms and show that equal-mass vessel fields can generate distinct functional masses and oxygenation levels.
Read the full abstract

Strongly Correlated Electrons

Bilinear magnon–exciton coupling in biased ferromagnetic electron–hole bilayers

The hybridization of magnons and excitons would combine magnetic and optical degrees of freedom in a single composite quasiparticle. Such a hybridization is however difficult to achieve, because of their inherent energy mismatch. We propose that in biased bipolar ferromagnetic electron–hole bilayers the excitons and magnons can be brought into resonance, with the exciton energy lowered through the voltage bias to match the magnon energies. We demonstrate the linear hybridization of magnons and spin-flip excitons in this regime, starting from the microscopic exchange interactions between electrons and localized magnetic moments. We show that further increasing the gate voltage softens the hybrid magnon–exciton mode and realizes a magnon–exciton condensate, which manifests in both the magnon and exciton sectors and is associated with spin-superfluid transport. Ferromagnetic electron–hole bilayers therefore provide a new platform for the study of composite magnon–exciton quasiparticles and the realization of spinful condensates and associated spin superfluidity.
Read the full abstract

Combinatorics

Soltés problem for the Kirchhoff index of a graph

We say that $v\in V(G)$ is a good vertex if the Kirchhoff index remains unchanged when $v$ is removed, i.e. $Kf(G)=Kf(G-v)$. In 1991, Šoltés studied the Wiener index of a graph and posed the problem of identifying graphs for which the removal of an arbitrary vertex preserves the Wiener index. In this paper, we explore a similar concept: identifying Kirchhoff Šoltés graphs, i.e. graphs in which all vertices are good vertices. We show that the cycle $C_5$ is a Kirchhoff Šoltés graph. We consider several relaxed versions of the Kirchhoff Šoltés problem, where the primary objective is to identify graphs containing at least one good vertex. One of them is the \textit{$\beta$-Kirchhoff Šoltés problem}, which seeks to find an infinite family of graphs in which the proportion of good vertices is at least $\beta$, with $\beta \in (0,1]$ being a specified rational number. Another one involves constructing infinite families of graphs where the proportion of good vertices increases and asymptotically approaches a given real number $\gamma\in (0,1]$ as the order of the graph grows. We demonstrate that both relaxed versions have infinitely many solutions. In particular, we prove the existence of infinitely many graphs for which the proportion $\beta$ of good vertices, $1/7\leq \beta<1/5$ tends to a certain irrational number. Furthermore, we prove the existence of infinitely many graphs with half good vertices, and for each $s\in\mathbb{N}$, we construct an infinite family of graphs whose proportion of good vertices tends to $\frac{s+1}{2s+1}$. These findings could be pivotal in addressing the original problem of determining whether there are additional solutions beyond $C_5$.
Read the full abstract

Analysis of PDEs

Blow-up or grow-up for the focusing 3D cubic NLS with a repulsive inverse-power potential at the mass–energy threshold

We consider the focusing cubic nonlinear Schrodinger equation with a repulsive inverse-power potential $V(x)=a|x|^{-\mu}$, where $a>0$ and $1<\mu\leq 2$. At the mass-energy threshold, Miao, Murphy, and Zheng, as well as Ardila, Hamano, and Ikeda, established sharp scattering results in the positive virial region. In this paper, we continue to study the dynamics in the negative virial region and show that, in each time direction, solutions either blow up in finite time or grow up.
Read the full abstract

Quantum Physics

Time domain non-linear quantum interferometry

Interferometry is a powerful method for studying many fundamental phenomena ranging from gravitational waves to anyon excitations. Besides loss of coherence in the interferometer paths, quantum mechanics sets an incompressible bound for linear interferometry arising from the quantum noise of the light beam, called the shot noise limit. Here, we implement a non-linear quantum interferometer that uses a microwave cavity and a magnetic field resilient anharmonic superconducting quantum circuit made of granular aluminum. The circuit enables interferometry beyond the shot noise limit. A direct application of our findings is the detection of dark matter (axions or dark photons), high-frequency gravitational waves or astronomical masers.
Read the full abstract

All authors C. Fruy, A. Théry, B. Hue, W. Legrand, L. Jarjat, J. Craquelin, M.R. Delbecq, A. Cottet, T. Kontos

Quantum Physics

Shot-based variational simulation of the toric code phase transition: Energy, global entanglement, and noise resilience

We investigate the capability of variational quantum circuits in detecting a topological phase transition within the toric code model subjected to a uniform magnetic field. The detection is carried out via shot-based computations of both the ground state energy and the global entanglement. We begin by employing well-established renormalization techniques to construct a quantum circuit that accurately represents the pure toric code under periodic boundary condition. This circuit is subsequently parameterized to enable a shot-based variational simulation of the ground state in the presence of the magnetic field. Our results demonstrate that even for the minimal lattice sizes and a finite number of measurement shots, the ground state energy exhibits a distinct singularity at the transition point which can be reliably approximated through finite-size scaling analysis. Furthermore, we introduce a secondary variational circuit designed to compute the global entanglement of the generated ground state. We observe that both global entanglement and its conditional counterpart display the expected qualitative behavior at the transition point, consistent with the topological nature of the phase transition. To evaluate the practical viability of our variational approach on real quantum hardware, we systematically examine the detrimental effects of gate noise on the analysis. We find that the topological phase exhibits a notable robustness against noise, manifesting as a linear decrease in the transition point with increasing noise strength.
Read the full abstract

Artificial Intelligence

How Should Reasoning Be Organized in a Transformer's Latent Space?

Continuous reasoning has emerged as a promising way to improve reasoning in large language models (LLMs). Yet we still lack a clear principle for deciding what a latent state should preserve. Reasoning by superposition shows that a single latent state can encode several search alternatives and expand them in parallel. We ask how those states should be weighted as reasoning proceeds. A natural choice is to preserve only the states active at the frontier step, since keeping every reached state appears to spread a limited hidden width too thin. We show that the opposite can hold. When later computation draws on several reached states, a cumulative state can guide attention correctly at a smaller hidden width than a frontier state that stores fewer states. At the same width, the cumulative state therefore keeps more intermediate states available for later reasoning. More generally, equal cumulative weights are optimal when future queries are unknown and remain close to the best task-specific weights when those queries are known. Experiments with two-layer and GPT-2 Transformers reproduce the predicted width advantage and show that unequal weights fail first on the states that receive the least weight. This suggests a important principle: keep reached states equally weighted, and restore equal weights as computation proceeds.
Read the full abstract

Systems and Control

Low-Ripple Modulation Strategy for a Photovoltaic-Based Triple-Port Hydrogen Production System

Among various production methods, hydrogen generation via electrolysis powered by renewable energy plays a key role in achieving large-scale green hydrogen production. The triple active bridge isolated DC-DC conversion system exhibits significant application potential in hydrogen production due to its advantages, such as high energy density, wide step-down ratio, and high reliability. However, the output current ripple at the hydrogen production port critically affects the efficiency of the electrolyzer and the hydrogen production rate. Existing studies have limited optimization effects on current ripple and struggle to achieve dynamic optimization, leading to constrained ripple suppression under dynamic operating conditions. To address this issue, this paper proposes a low-ripple modulation strategy based on coordinated optimization of inner and outer phase-shift angles for multi-port power conversion systems in renewable energy hydrogen production. By establishing an accurate mathematical model, the optimal phase-shift angle combination under minimal current ripple conditions is derived. An improved differential evolution algorithm with adaptive parameter strategy is employed to achieve global optimization under dynamic conditions. Simulation and experimental results demonstrate that the proposed strategy effectively suppresses current ripple, providing an efficient and reliable solution for hydrogen production from fluctuating renewable energy sources.
Read the full abstract

All authors Shiqi Zhang, Ziang Jiao, Jiaxin Su, Ning Wang, Zheng Li, Xiaoqiang Guo, Changchun Hua

Computation and Language

Inside VLM Chart Reading: Tracing Value Reading from Vertical Bar Charts Across Space and Depth

Vision-language models (VLMs) can answer chart questions accurately, but output accuracy alone provides limited insight into the internal representations supporting exact value reading. We study value reading from vertical bar charts through controlled counterfactual activation patching in Qwen2.5-VL-7B-Instruct and InternVL3.5-8B. Our study connects three analyses. First, under one-tick bar-height counterfactuals, patching states selected by the changed-height region produces greater restoration of the original answer preference than patching states selected by the invariant bar body. Across language-model depth, restoration from legend- and series-related positions declines earlier than restoration from bar-geometry and axis-scale positions. Second, we identify middle-layer states at series-related prompt positions that can restore answer preference under legend swaps. Resetting these states selectively weakens restoration from an earlier visual-legend patch, demonstrating their causal role in the downstream effects of that intervention. Third, a factorial analysis shows that both models can combine geometry and scale states from separate donors to favor the combined target answer. Together, these results provide causal evidence about the internal states supporting exact bar-value reading.
Read the full abstract

Quantum Physics

Room-Temperature Storage of Entanglement in a Silicon Carbide Quantum Node

Robust entanglement at room temperature is a central challenge for solid-state quantum information processing and quantum-enhanced sensing. Here we demonstrate room-temperature storage of entanglement in a silicon carbide (SiC) quantum node by coherently transferring an electron-nuclear entangled state onto long-lived nuclear-spin memory qubits. Using a shallow single color center in 4H-SiC, conventionally denoted PL6, we realize a fully addressable three-qubit register composed of one electron-spin processor and two strongly coupled $^{29}$Si nuclear-spin memory qubits. This platform enables the deterministic generation of high-fidelity entangled states, including a nuclear-spin Bell state with a fidelity of $94 \pm 2\%$ and a three-qubit Greenberger-Horne-Zeilinger (GHZ)-type state with a fidelity of $89 \pm 4\%$. By implementing a SWAP-gate protocol in the strong hyperfine-coupling regime, the electron-nuclear entanglement is transferred to the nuclear-spin memory with a fidelity of $92.5 \pm 2.5\%$, extending the entanglement lifetime by a factor of 240. We further confirm the generality of this approach in an additional heterogeneous $^{29}$Si-$^{13}$C nuclear-spin register and, through a statistical survey of 200 single PL6 centers, show that multi-nuclear-spin registers occur naturally with probabilities above 10%. These results position shallow SiC color centers as a powerful platform for entanglement-assisted quantum sensing and scalable room-temperature quantum technologies.
Read the full abstract

All authors Shuo Ren, Rui-Jian Liang, Qi-Cheng Hu, Zhen-Xuan He, Ji-Yang Zhou, Wu-Xi Lin, Zhi-He Hao, Tao Tu, Jin-Shi Xu, Chuan-Feng Li, Guang-Can Guo