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

Friday 11 September

Statistical Computation

Wasserstein mixing of a systematic-scan random rotation sampler

We study the mixing time of a systematic-scan analogue of Kac's walk that was proposed as a fast surrogate for Haar-distributed orthogonal matrices in randomized high-dimensional algorithms and was conjectured to approach Haar measure after only logarithmically many sweeps. We show that this conjectured speed-up does not occur for convergence of the full matrix law to Haar measure in Frobenius Wasserstein distance. At fixed normalized accuracy, the mixing time lies between order $n/\log n$ and order $n$ sweeps; at fixed absolute Frobenius accuracy, the corresponding bounds are between order $n$ and order $n\log n$. More strongly, below the scale $n/\log n$, the normalized Wasserstein distance remains asymptotically at its extremal value. We also show that the output law is singular with respect to Haar measure for fewer than $n/2$ sweeps. Thus the sampler may provide effective application-specific randomization without exhibiting the much faster full-Haar mixing.
Read the full abstract

Optimization and Control

Formation control from the generic combinatorial viewpoint: edge dynamics and directed sensing

We develop a geometric framework for distance-based formation control that separates the evolution of inter-agent distances from its realization by compatible node motions, reducing the stability problem to the edge space. We show that local exponential convergence of the edge dynamics implies local exponential convergence of the formation, and that stability is certified by spectral properties of a linear edge operator. We introduce a hierarchy of generic spectral properties — weak admissibility, admissibility, and strong admissibility — that provide necessary conditions for local exponential stability. Specializing to directed sensing, we obtain a necessary and sufficient spectral condition for local stability at an arbitrary target, together with a quadratic sufficient certificate. These conditions reveal that stability depends jointly on the graph orientation and target geometry, and show that persistence is neither necessary nor sufficient for local convergence. We show that every generically rigid graph admits an admissible orientation and, for acyclic orientations, we give an exact combinatorial characterization of admissibility. Finally, the quadratic certificate leads to a semidefinite program for synthesizing stabilizing edge gains.
Read the full abstract

All authors Louis Theran, Daniel Zelazo, Sean Dewar, Bernd Schulze

Artificial Intelligence

Vibe Patenting: Evaluating LLM Judges for Professional Patent-Drafting Agents

LLM judges are increasingly used to evaluate and improve AI-generated outputs, yet their reliability for complex professional work remains unclear. We study this problem through Vibe Patenting, an end-to-end patent-drafting testbed for AI-agent evaluation. A separately-invoked LLM judge evaluates generated patent drafts and provides structured feedback for iterative revision. Across multiple inventions and drafting-agent configurations, judge-guided revision consistently improves judge-assessed quality, while unguided revision tends to saturate. Notably, iterative judge feedback enables a low-reasoning agent to approach the performance of a substantially more expensive high-reasoning agent. Stronger models and increased reasoning generally improve judge-assessed drafting quality, while domain-specific agentic workflows provide further gains. We validate the judge against independent evaluation by a professional patent attorney and find meaningful but strongly metric-dependent agreement and systematic calibration differences. These results highlight both the utility and limitations of LLM judges as evaluators and optimization signals for complex professional workflows.
Read the full abstract

All authors Toshiaki Koike-Akino, Vladislav Blaykhman, Ye Wang, Jing Liu, Gene V. Vinokur

Earth and Planetary Astrophysics

The core-mantle mode of gravitational oscillation

We present the characteristics of a mode of axial oscillation between Earth's mantle, fluid core and solid inner core that has not been previously reported. The mode involves a quasi-rigid rotation of the fluid core outside the tangent cylinder (TC) exchanging its angular momentum with the mantle via a three step process. First, by a magnetic torque with the fluid inside the TC; second by a magnetic torque between the latter and the inner core; and finally by a gravitational torque between the inner core and mantle. Although the gravitational torque is purely between the mantle and inner core, the mode involves an oscillation of the whole of the core, and we refer to it as the core-mantle gravitational (CMG) mode. This form of gravitational oscillation occurs when the magnetic field within the core is sufficiently strong that the propagation time of Alfvén waves is shorter than the mode period, which is the case for Earth. We show how the period, quality factor $Q$ and structure of the CMG mode depends on the strength of the gravitational torque and viscous relaxation time $\tau_i$ of the inner core. For Earth, the CMG mode period should be in the range of 40 to 100 years, but viscous relaxation of the inner core likely implies a small $Q$, below 1 if $\tau_i<10$ years. Our results suggest that the CMG mode may act to amplify resonantly, though only modestly, multi-decadal changes in the length of day driven by zonal accelerations in the fluid core.
Read the full abstract

Hardware Architecture

Arborist: Algorithm-Hardware Co-Design for Fast and Efficient Motion Planning

Real-time motion planning must execute under strict latency and energy constraints on resource-limited platforms. This paper presents Arborist, an algorithm-hardware co-design framework that accelerates Fast Marching Tree (FMT*) motion planning through synergistic algorithmic and architectural innovations. At the algorithm level, Arborist introduces safe multi-tree expansion with global competition management and cross-tree optimality checking to expose inter-tree parallelism while preserving path quality. At the architecture level, our proposed ArboristAccel integrates an Arborist Toolbox that provides hardware support for inter-tree parallelism; a KD-tree Subsystem for high-throughput near-neighbor search; a Parallel-Query Retrieval module that balances multi-bank memory accesses to alleviate system-level memory bandwidth bottlenecks; and a Look-ahead Planning Engine that exploits intra-tree parallelism while preserving path quality by in-order commit. Synthesized in 16 nm and operating at 500 MHz, ArboristAccel sustains more than 1000x speedup over CPU across all workloads, peaking at 2800x on 2D-Maze, and delivers geomean improvements of 8.8x speedup, 4.6x area efficiency, and 3.1x power efficiency over an FMT* ASIC baseline. The corresponding factors relative to a state-of-the-art RRT*-based MOPED accelerator are 8.6x, 1.0x, and 1.8x, at comparable path cost. To the best of our knowledge, ArboristAccel is the first dedicated hardware accelerator for FMT*-based motion planning.
Read the full abstract

All authors Yaotian Liu, Lingyi Huang, Zishen Wan, Bo Yuan, Cheng Tan, Jeff Zhang

Discrete Mathematics

A Four-Connected Graph without a Legal System

In a 2021 paper, Jankiewicz, Norin, and Wise asked whether there exists a finite $4$-connected graph of girth at least four and nonnegative Charney–Davis curvature such that no $4$-connected ordinary subgraph admits a legal system. We construct such a graph by starting from the hexagonal prism and attaching three $K_{3,4}$-based caps along pairwise disjoint induced $4$-cycles. The key structural input is a restriction theorem showing that a legal system on an induced-$4$-cycle amalgam restricts to each side, so the obstruction carried by the negatively curved prism survives the attachments. The resulting $33$-vertex graph is $4$-regular and $4$-connected, has girth four and Charney–Davis curvature one, and, by $4$-regularity, is its own unique $4$-connected ordinary subgraph.
Read the full abstract

Quantum Physics

High quantum local differential privacy breaks entanglement

Differential privacy provides a mathematical framework for guaranteeing privacy for sensitive data. In quantum information processing, the interaction of privacy constraints with quantum resources such as entanglement remains a question of interest. Given that the utility of many protocols, and often the presence of a quantum advantage, relies on quantum resources such as entanglement, it is crucial to understand when a privacy requirement for a quantum channel is compatible with the channel's ability to preserve entanglement. We study this question for quantum local differential privacy (QLDP). Our main result shows that every $\varepsilon$-QLDP channel with a $d$-dimensional input is entanglement-breaking whenever $\varepsilon\leq\log\frac{d}{d-1}$. We also prove an approximate version for $(\varepsilon,\delta)$-QLDP, where channels in the same high-privacy regime are close in diamond norm to an entanglement-breaking channel. We further prove a composition result for a collection of private quantum channels having entangled inputs and global measurements in the high-privacy regime. Finally, we apply our results to private quantum learning theory. We prove that any learning protocol using arbitrary quantum memory on copies of the output of an entanglement-breaking channel can be simulated by a protocol that measures the corresponding unprocessed input copies one at a time while storing only classical information. Combining this result with our high-privacy entanglement-breaking theorem, we show that under sufficiently private local noise, a learning protocol with quantum memory for purity testing and bipartite product testing is subject to the sample complexity lower bounds for protocols with single-copy measurements on the noiseless tasks. We also obtain stronger sample complexity lower bounds when a single highly private channel acts on the entire multipartite input.
Read the full abstract

Combinatorics

Prescribed Turán sign patterns for inverse Kazhdan-Lusztig polynomials of matroids

Gao and Xie conjectured that the coefficients of the ordinary, unnormalized inverse Kazhdan-Lusztig polynomial of every matroid are log-concave with no internal zeros. We give counterexamples and prove locality and composition formulas for deletions of projective charts. For every finite field and every nonempty finite set S contained in {2, 3, ...}, these formulas yield simple 3-connected matroids whose negative internal Turán determinants occur exactly at S. Their coefficients are positive and strictly decreasing. The polynomial degree can be prescribed above max S, and all selected Turán ratios can tend to zero simultaneously. The universal first-index inequality remains open.
Read the full abstract

Probability

Coloured Epidemic Models: Functional Law of Large Numbers and Propagation of Chaos

In this paper, we study a stochastic Susceptible-Infected-Removed (SIR) model where the infection and the recovery rates depend on individual covariates for susceptibility and infectiousness of the infector and the infectee. Such models allow explicit nonlinearity in the incidence term. They are also important from a practical perspective, as they allow for the incorporation of individual heterogeneity into the epidemic process. Statistical estimates for crucial epidemiological parameters, such as the basic reproduction number, herd immunity threshold, could be vastly different, and even biased, when the population heterogeneity is ignored in the mathematical model. We describe our epidemic model as an Interacting Particle System (IPS) of Stochastic Differential Equations (SDEs) driven by Poisson Random Measures. Our main mathematical contributions are a Functional Law of Large Numbers (FLLN), which approximates the empirical random measure of the IPS by means of a deterministic measure-valued function, and the propagation of chaos phenomenon, which establishes asymptotic independence of the particles as the population size goes to infinity with an explicit construction of McKean–Vlasov type Kac's “nonlinear process”. We also briefly mention how the propagation of chaos phenomenon leads to a product-form likelihood function, which forms the basis of the so-called Dynamic Survival Analysis (DSA) method for parameter inference based on sparse data.
Read the full abstract

All authors Kushankur Dutta, Olga Izyumtseva, Wasiur R. KhudaBukhsh, Grzegorz A. Rempała

Machine Learning

Specification Oracles

Specifications face a basic tradeoff: leave details out, and important questions go unanswered; record every detail separately, and the specification becomes large and prolix. We investigate whether a language model can serve as a compact, living specification oracle by learning facts about a target and answering questions about it directly. We compare two ways of storing the learned facts: external text notes and changes to the model's weights. Across four families of 596-fact worlds and two Qwen2.5 model sizes, weight-only oracles benefited substantially more from structure: with the 7B model, their accuracy integrated across storage capacities was 18.5 percentage points higher on structured than unstructured worlds, compared with 1.1 points for note-sheet oracles. This advantage came at a substantial storage cost, with the smallest adapter requiring approximately 175 KiB compared with a maximum note budget of 16 KiB. Adapted weights therefore exploited latent structure more successfully, while external notes required substantially less object-specific storage.
Read the full abstract

Analysis of PDEs

Crossover asymptotics and a sharp confinement rate for the viscous Burgers equation

We study the one-dimensional viscous Burgers equation on $(0,L)$ with the conservative boundary conditions $u_x=-u^2$. On the half-line, solutions approach a nonlinear self-similar profile, whereas on a bounded interval they converge to a nonconstant equilibrium. We describe explicitly how the dynamics passes between these states at the critical diffusive scale $t=cL^2$. For compactly supported initial data of mass $M$, the rescaled solution converges as $L\to\infty$ to an explicit crossover profile $\Phi_c$. In similarity variables, $\Phi_c$ converges to the half-line profile $f_M$ as $c\downarrow0$; after rescaling to domain variables, it converges to the interval equilibrium as $c\to\infty$. We also determine the sharp onset of confinement: \[ \lim_{c\downarrow0}-c\log|\Phi_c(0)-f_M(0)|=1\qquad(M\ne0). \] Moreover, on compact sets in similarity variables, the interval and half-line solutions differ in $C^k$ by at most $C_{k,\varepsilon}\exp(-(1-\varepsilon)L^2/t)$, uniformly in $L$, and the exponential constant is optimal. Thus we identify not only the transition scale $L^2$, but also the profile governing the crossover and the sharp rate at which the remote boundary becomes visible. We finally discuss the conclusions that persist for space-dependent diffusivity and illustrate the three asymptotic regimes numerically.
Read the full abstract

Computation and Language

CVSS-X: A Multilingual Speech-to-Speech Translation Corpus for 28 Languages

We introduce CVSS-X, a large-scale synthetic speech-to-speech translation corpus that extends CVSS by reversing the translation direction. While CVSS translates from 21 languages into English, CVSS-X enables translation from English into 28 target languages spanning 12 language families. The corpus comprises approximately 240,000 parallel speech pairs per language, totaling over 16,000 hours, eight times larger than CVSS. We provide two variants: CVSS-X-C with two canonical voices per language, and CVSS-X-T with cross-lingual voice cloning, both fully generated. Evaluation shows comparable translation quality to CVSS with consistent performance across typologically diverse languages. Combined with CVSS, this enables research on bidirectional and multilingual speech-to-speech translation. The code is available at https://github.com/ErmisAI/XVSS-X and the dataset under CC-BY-NC 4.0 license at https://huggingface.co/datasets/lgris/XVSS-X.
Read the full abstract

All authors Lucas Rafael Stefanel Gris, Alef Iury Siqueira Ferreira, Frederico Santos de Oliveira, Augusto Seben da Rosa, Alexandre Costa Ferro Filho, Arlindo Rodrigues Galvão Filho, Anderson da Silva Soares

Combinatorics

Lieb's Permanental Dominance Conjecture for Ordinary Immanants through Order Fifteen

Pate proved ordinary irreducible-immanant permanental dominance through order $13$ and identified $(4,4,3,3)$ as the sole remaining order-$14$ case, with $(5,4,3,3)$ and $(3^5)$ forming the order-$15$ frontier. These three cases are settled here; consequently $d_\lambda(A)/f^\lambda\le \operatorname{per}(A)$ for every partition $\lambda\vdash n$ with $n\le15$ and every complex Hermitian positive-semidefinite matrix $A$. The argument also yields results beyond this finite frontier: an exact four-term bridge for $(4,4,3,3)$, the uniform family $(m,4,3,3)$, a two-parameter family $(a,b,3,3)$ for $a\ge b\ge4$ and $5a\ge8b$, and a long-first-row criterion for arbitrary fixed tails. These results arise from explicit specializations of Pate's $W$-function positivity framework using partial swaps, Young projectors, Pieri–content identities, and branching data. For $(3^5)$, an exact Farkas certificate shows that the central-projector partial-swap cone is insufficient; a branching-refined one-swap construction escapes this obstruction and yields a positive $106+19$-witness certificate. Boundary-compression and node-moving results further describe the reach and limitations of the local-filter method. All finite certificates are checked by exact integer or rational arithmetic and are supplied as ancillary material. The order-$14$ bridge is additionally formalized and kernel-checked in Lean 4 for all complex Hermitian positive-semidefinite matrices, including the exact coefficient normalization and the deduction of $(4,4,3,3)$ permanental dominance from four explicitly stated Pate inequalities.
Read the full abstract

Strongly Correlated Electrons

Tripartite entanglement in chaotic eigenstates

It is now well-established that classical statistical mechanics emerges from the entanglement structure of quantum chaotic systems, as quantified by the (subsystem) eigenstate thermalization hypothesis (ETH). While this statement rests on the well-studied bipartite entanglement of the eigenstates of such systems, recent years have shown that many-body states are often characterized by multipartite entanglement, raising the question: what are the universal multipartite entanglement features of chaotic eigenstates? We answer this question in part by studying a tripartite generalization of the ergodic bipartition ansatz through the lens of two tripartite entanglement probes, the Rényi tripartite multi-entropy and the Rényi reflected entropy. We derive leading-order analytical expressions for the ensemble averages of both quantities using a saddle-point analysis. Beyond these measures, we obtain the average of the Rényi-2 reflected density matrix, $\rho_{AA^*}^{(2)}$,which controls the Rényi-2 reflected entropy. We show that $\rho_{AA^*}^{(2)}$ in a chaotic eigenstate is well-described by this ensemble average which, unlike an ordinary reduced density matrix, takes three qualitatively distinct forms across subsystem-size regimes, which may be viewed as a tripartite generalization of subsystem ETH. We benchmark our analytical predictions against exact diagonalization of the mixed-field Ising model.
Read the full abstract

Soft Condensed Matter

Biaxial nematics and nematic-nematic demixing in polydisperse mixtures of hard board-like particle fluids

We study the bulk phase behavior of a polydisperse liquid crystal fluid made of biaxial boards with restricted orientations using both fundamental measure theory and Monte Carlo computer simulations. The continuous polydispersity is included in the intermediate particle length of the boards via a truncated Schulz distribution. By calculating several phase diagrams across a range of polydispersity coefficients, we find that polydispersity (i) enhances demixing between two uniaxial nematic phases and (ii) expands the stability region of the biaxial phase. Although metastable with respect to non-uniform phases, we identify certain mixtures exhibiting two-phase coexistence paths involving uniaxial-uniaxial and uniaxial-biaxial phase separations. Whether these paths can be completed depends on the precise shape of the parent distribution function. Monte Carlo simulations performed for a representative case exhibit the same phase diagram topology predicted by theory, thereby validating the theoretical approach. The main difference between both approaches lies in quantitative agreement, with simulated phase transitions systematically occurring at higher packing fractions than those predicted theoretically. Our combined theoretical and simulation results may prove relevant to the design and interpretation of sedimentation experiments on colloidal suspensions of polydisperse anisotropic particles.
Read the full abstract

Image and Video Processing

Tile-Based Phase Unwrapping with Reliability-Ordered Joining and Residual-Weighted Multipath Averaging

Phase unwrapping recovers a continuous phase map from measurements known only modulo 2{\pi}, in the presence of noise and discontinuities. We propose a tile-based method that combines local spectral reconstruction, reliability-ordered tile joining, and residual-weighted multipath averaging. Each tile is reconstructed by a discrete cosine transform (DCT) least-squares solve, and robust measurements along tile borders guide their joining along a spanning tree. We repeat this reconstruction on shifted tile grids and on rotated and reflected copies of the image, which changes tile borders and joining paths. A residual-weighted average combines the resulting reconstructions, which are computed in parallel and averaged on the fly without being stored. We also introduce the corrected circular rewrapping error, a congruence measure: after removing a global phase offset, it measures how far the rewrapped reconstruction is from the observed phase. It applies to least-squares and path-based methods, even without an unwrapped reference. Used with reference-based errors, it separates agreement with the measurement from recovery of the correct number of cycles. Experiments on radar-interferometric, magnetic-resonance, holographic, and constructed phase images compare our method with seven reference algorithms. Among the methods that complete every test image, ours has the lowest mean fraction of pixels with an error above {\pi}. On the images all methods complete, it lies on several speed-accuracy Pareto fronts. Error maps also reveal isolated cycle errors that average scores or congruence alone can hide.
Read the full abstract

Superconductivity

Topological superconductivity in an altermagnet-superconductor heterostructure

Fully gapped spin-triplet superconductors offer a natural setting for topological edge transport and Majorana excitations. Here we propose a heterostructure that interfaces altermagnetic monolayer V$_2$Se$_2$O with a conventional $s$-wave superconductor. Starting from an ab-initio description of V$_2$Se$_2$O, we construct a low-energy model of its spin-split Fermi surface and study the superconducting instability of the coupled stack. We find that the momentum-dependent exchange splitting of the altermagnet strongly favors equal-spin pairing and converts the proximity-induced order into a fully gapped $p$-wave state. Over a broad range of interface parameters, the resulting Bogoliubov-de Gennes bands carry a nonzero Chern number, and slab spectra exhibit chiral Majorana edge modes traversing the bulk gap. We further quantify the inverse proximity effect and show how the induced gap evolves across a finite multilayer stack. These results establish a route to topological triplet superconductivity by interfacing a conventional superconductor with an experimentally available altermagnetic semiconductor, and provide practical guidance for interface fabrication and experimental detection.
Read the full abstract

All authors Michael Liudeng, Hrishikesh Patel, Marcel Franz, Niclas Heinsdorf

Cryptography and Security

Principled Detection of Coordinated Manipulation from Aggregate Distortion and Account Reuse

Coordinated manipulation is collective: plausible accounts can jointly distort ratings, rankings, and engagement. Existing defenses primarily construct evidence from identities, graphs, content, or co-activity. We introduce an aggregate-first evidence layer that treats distortion of a context-level outcome distribution as the primary evidence object. The engine observes only a histogram, count, resolution, and reference distribution; identities are withheld until interval evidence is fixed. Because raw discrepancies have positive finite-sample expectation, we subtract a matched null expectation to obtain signed evidence and account for reference uncertainty. Participation logs then accumulate these fixed increments across accounts. We characterize matched-exposure divergence, bound self-influence, establish finite-horizon separation, and derive an exact linear reuse law for paired contexts. We evaluate the mechanism with controlled rotation experiments and paired counterfactual interventions on historical Amazon review streams. Historical reviews provide the behavioral background; synthetic identities provide known coalition membership, and exact clean twins provide counterfactual controls. In a fixed-attack sweep against historical non-donor comparison accounts, reassigning the same manipulated events across identities with increasing reuse raises account-score ROC-AUC from 0.500 to 0.797. With activity- and exposure-matched clean twins, frequency is at chance while counterfactual attribution achieves ROC-AUC 0.744. Under a mean-preserving shape intervention, Wasserstein-1 and Jensen-Shannon evidence achieve ROC-AUC 0.909 and 0.967, while frequency and mean-based attribution remain at chance. Aggregate evidence complements repeated co-activity, improving mixed-mechanism ROC-AUC from 0.750 to 0.874 with a simple untrained combination.
Read the full abstract

Artificial Intelligence

Generalized Agent Iteration: One Formal Framework for Iterative Policy Improvement and Recursive Self-Improvement

When we speak of recursive self-improvement (RSI), are we speaking of a phenomenon, a mechanism, or a prospect? Towards autonomous and evolving intelligence, RSI is being claimed at many scales, while no single framework that formally describes these emerging instances exists. Its counterpart in the classical realm, iterative policy improvement, is characterized by generalized policy iteration (GPI), a framework of broad applicability with well-understood theoretical properties, but only where the update principle and the evaluation base lie outside the agent. In this paper, we propose Generalized Agent Iteration (GAI), a formal framework that describes iterative policy improvement and RSI as two cases of a single learning paradigm. GAI defines the agent as a configuration of modifiable components within a system and models the learning process as a cycle of agent evaluation and agent improvement. Two pivotal dials then distinguish the instances, in particular the systems that self-identify as RSI in recent work: whether the improving mechanism is part of the agent and whether the standard it is measured against is grounded outside it. The former dial delineates the boundary between GPI and RSI, and the latter determines a system's polarity as anchored, goal drift, or fully self-referential. Moreover, we use these coordinates to place existing systems on the same two axes and make the defects of recursive self-improvement statable one condition at a time. We see this paper as a first step toward exploring a formal characterization of RSI that rests on the classical account, makes existing systems comparable, and provides a principled basis for analyzing and designing new ones.
Read the full abstract

All authors Hongyao Tang, Yi Ma, Pengyi Li, Yifu Yuan

Signal Processing

Krum-Inspired Central Teacher Selection and Residual Channel Bottlenecks for Efficient DeepSC

Deploying transformer-based semantic communication models on edge devices requires compression that preserves semantic fidelity under channel variability. We study a compressed deep semantic communication (DeepSC) student trained by multi-teacher knowledge distillation and combine two ideas: (i) a residual channel bottleneck that splits the transmitted representation into a base stream and a residual stream with unequal power allocation, and (ii) a Krum-inspired, medoid-style centrality criterion that selects a single central teacher from a five-model ensemble, applied at the logit and intermediate-feature levels and combined with a feature-dominant distillation loss. On the EuroParl benchmark, a two-layer student recovers about 98% of the four-layer teacher's bilingual evaluation understudy (BLEU)-1 under additive white Gaussian noise, with 93% of its BLEU-4 and 96% of its sentence-BERT (SBERT) score, and about 89%, 77%, and 87% of the teacher's BLEU-1, BLEU-4, and SBERT, respectively, under Rayleigh fading, while reducing non-embedding parameters by 1.33x and single-teacher inference latency by 1.79x (9.0x relative to a five-teacher ensemble used here as an upper-bound reference, not a deployment baseline). Controlled ablations over three seeds indicate complementary contributions of about +8.4% BLEU-1 (+9.6% SBERT) from the residual bottleneck and +3.0% (+2.9% SBERT) from centrality-based selection over mean aggregation. A decoder-mode ablation shows that the base stream alone recovers about two-thirds of the full BLEU-1 while the residual stream alone collapses, supporting the role of the residual as a refinement on top of the base.
Read the full abstract

All authors Rami Eid, Mostafa Jammoul, Omar Kaaki, Maria Slim, Mariette Awad, Hadi Sarieddeen

Astrophysics of Galaxies

Mapping parameters of idealised hydrodynamic galaxy simulations to bar properties: a case study with the RAMSES code

Cosmological simulations with high spatial resolution often struggle to reproduce the observed fraction of galactic bars across redshift and may produce bars that are too short. To investigate these issues from a purely Newtonian dynamical perspective, we run a grid of idealised hydrodynamic simulations of disc galaxies with stellar masses of $\sim10^{10},{\rm M}_\odot$ using RAMSES. Without cooling, star formation, or feedback, we vary the resolution, gas mass, stellar velocity dispersion, bulge mass, halo mass, and concentration to identify the conditions that inhibit bar formation. We test whether initially axisymmetric discs form bars over the typical time elapsed between $z\sim1$ and $z\sim0.2$. We find that most diagnostics proposed in the literature are too simplistic to reliably predict bar formation. However, a region of parameter space that strongly inhibits bar formation is identified: a high Romeo-Falstad stability parameter, with a threshold that decreases quadratically with bulge mass, combined with a high generalised Efstathiou-Lake-Negroponte (ELN) parameter that accounts for the bulge. Lower gas fractions and larger numbers of dark matter particles also tend to reduce the bar growth rate. We argue that if a bar is destroyed, for example by bulge formation, after the galaxy enters this bar-inhibiting region, re-formation may be difficult. This suggests that a modern version of the angular momentum catastrophe may persist in large-volume cosmological simulations. Finally, we confirm that only baryon-dominated discs lying away from the stellar-to-halo-mass relation expected from abundance matching can form bars sufficiently large relative to their corotation radius.
Read the full abstract

All authors Srikanth T. Nagesh, Benoit Famaey, Jonathan Freundlich, Yves Revaz, Giacomo Monari, Arnaud Siebert, Rodrigo Ibata

Functional Analysis

The essential norm, block sizes, and some Generalized Hilbert operators on lp

We examine the generalized Hilbert (matrix) operators $$ H_{g,\gamma} : (a_n) \mapsto \sum_{n=1}^{\infty} \bigg(\frac{k}{n}\bigg)^{\gamma} \frac{g_k a_n}{n+k} $$ on the $\ell^p$ spaces, $1<p<\infty$, where $g=(g_n)$ is a sequence and $-1/p<\gamma <1-1/p$. Given a partition of the natural numbers $\bigcup_j I_j = \mathbb N$, we encode $g$ to our investigation via its $\ell^p$-means $G_j$ over the sets $I_j$. We prove that if $I_j$ increases exponentially, the data $(G_j)$ characterizes boundedness, but is insufficient to determine the exact value of the essential norm. Nonetheless, if the growth rate of $(I_j)$ is subexponential, the corresponding data $(G_j)$, given that the tail converge, is sufficient to determine the exact value, in which case we also calculate it. Moreover, it is shown that the boundedness of $(G_j)$ is sufficient, but not necessary (and necessary, but not sufficient) to ensure $H_{g,\gamma}$ is bounded if the underlying partition $(I_j)$ increases subexponentially (and superexponentially, respectively). Using the dual operator, we also prove that the essential norm of $H_{g,\gamma}$ is comparable with $\limsup_j G(j)$ when $(I_j)$ grows exponentially.
Read the full abstract

Computational Finance

Diffusion models for dynamic volatility surface generation and data-driven hedging

We develop a diffusion-model framework for dynamic implied-volatility surface generation and evaluate its economic usefulness through data-driven hedging. The framework consists of two models. AD-Seq-Vol jointly learns the conditional evolution of the underlying asset return and the high-dimensional implied-volatility surface, generating adapted multi-period scenarios by sequentially updating the realized market history. AD-Seq-Vol-FT further incorporates option-market structure through post-training penalties for violations of static no-arbitrage conditions. Using daily SPX option data from 2000 to 2023, we show that the proposed models generate coherent surface trajectories while capturing both cross-sectional and temporal dependence. AD-Seq-Vol produces fewer and less severe static-arbitrage violations than the training data and the GAN-based benchmark, while AD-Seq-Vol-FT reduces these violations to nearly zero. We then integrate the generated conditional scenarios into an optimization-based hedging framework. Compared to a range of classical and data-driven benchmarks, the diffusion-based hedges maintain tracking errors near zero, substantially reduce tail risk, and exhibit particularly stable performance during the COVID-19 market disruption. These results establish Adaptive Sequential Diffusion Models as a promising class of market-consistent and economically useful financial scenario generators. Our code is available at: https://github.com/yinbinhan/volatility-surface-simulation.
Read the full abstract

All authors Yinbin Han, Jack Yuxiang Zhang, Manuel Torres, Fernando Acero, Renyuan Xu

Logic in Computer Science

Perturbation equivalence in labelled Markov chains

Behavioural equivalences, such as language equivalence and probabilistic bisimilarity, are fundamental techniques for reducing the size of probabilistic models. However, these equivalences are sensitive to the precise values of transition probabilities, making them unsuitable in applications where probabilities are subject to approximation. Motivated by settings in which the support graph of a labelled Markov chain is known but the transition probabilities are uncertain, we study robust variants of these equivalences. We introduce universal (perturbation) equivalence, which captures a variant of equivalence that is resilient to all perturbations of transition probabilities: two states or distributions are universally equivalent if they remain equivalent under every assignment of transition probabilities consistent with the support graph. We also consider the dual notion of existential (perturbation) equivalence, which holds whenever there exists an assignment of transition probabilities that yields equivalence. We establish that, for states, universal language equivalence coincides with universal probabilistic bisimilarity and develop a characterisation that yields a polynomial-time partition refinement algorithm. We implement the algorithm and demonstrate experimentally that it is effective as a technique for robust model reduction. We further show that universal language equivalence for distributions is closely related to the state case, and prove NL-completeness of deciding universal equivalence for both states and distributions. We prove that, for states, existential language equivalence coincides with existential probabilistic bisimilarity and deterministic witness transition functions always suffice, leading to an NP-completeness result. In contrast, we show that existential language equivalence for distributions is complete for the existential theory of the reals.
Read the full abstract

All authors Syyeda Zainab Fatmi, Stefan Kiefer, James C. A. Main, David Parker

Analysis of PDEs

Local rigidity of conformally Euclidean metrics for the anisotropic Calderón problem

We prove local rigidity of every smooth conformally Euclidean metric for the anisotropic Calderón problem on smooth compact domains $M\subset\mathbb R^n$, $n\ge3$. Any smooth Riemannian metric $g$ sufficiently close to a fixed background $g_0=e^{2c}\mathsf e$ in $H^s(M)$, with integer $s>n/2+1$, having the same Dirichlet-to-Neumann map satisfies $g=\Phi^*g_0$ for a smooth diffeomorphism $\Phi$ fixing the boundary pointwise.
Read the full abstract

Superconductivity

Superconducting Diode Effect in Josephson $\varphi_0$ Junction

We investigate the superconducting diode effect in a Josephson junction with a ferromagnetic weak link in the presence of Rashba spin-orbit coupling. While the standard $\varphi_0$ junction model is valid only for weak exchange fields $h \lesssim T_c$, realistic ferromagnetic barriers typically operate in the regime $h \gg T_c$, where the conventional Ginzburg-Landau gradient expansion becomes insufficient and requires the inclusion of higher order gradient terms to ensure stability of the free energy. By extending the free energy with such higher order term, we demonstrate that the linear gradient coupling alone leads to finite momentum Cooper pairing while leaving the critical current symmetry unchanged. The interplay between this linear term and the higher order gradient one leads to the Josephson diode effect.
Read the full abstract

Materials Science

Visualizing impurity-driven scattering phase textures in EuCd2As2

Understanding how disorder modifies electronic states in magnetic semiconductors is important for controlling spin-dependent transport and topological responses. Here we use scanning tunneling microscopy to visualize scattering phase textures in EuCd2As2. By isolating a single surface wavevector we reconstruct spatial phase maps of the local density of states and identify phase dislocations characterized by 2pi winding around impurity sites. These phase singularities emerge systematically within charge puddles generated by Eu interstitials and their positions evolve with bias voltage. We show that their spatial structure is consistent with interference between multiple scattering channels, including contributions from spin-orbit coupling. We provide a model which reproduces phase dislocations and relates the decay of the phase gradient to the relative strength of spin-orbit and scalar scattering. Our results establish a route to access the phase of electronic scattering in real space and study the role of local disorder and spin-orbit interactions in shaping electronic states in quantum materials.
Read the full abstract

All authors Raquel Sánchez-Barquilla, Rafael Pineda Medina, Pablo García Talavera, Adrian Valadkhani, Edwin Herrera, Brinda Kuthanazhi, Lin-Lin Wang, Sergey L. Bud'ko, Paul C. Canfield, Roser Valentí, Isabel Guillamón, William J. Herrera, Alfredo Levy Yeyati, Hermann Suderow

Computer Vision

ConeGaussian: Anti-Aliased Gaussian Ray-Tracing for Generic Central Cameras

In rendering, a camera is a sampling operator that maps each finite pixel to a bundle of rays. Different camera models change the geometry of this bundle, thus making a unified and faithful rendering formulation challenging. Consequently, Gaussian ray tracing supports generic cameras (with optical center) through their inverse ray mappings, yet typically reduces every pixel to a single center ray. This ignores the camera-dependent pixel footprint, causing aliasing under minification, while unconstrained Gaussians expose unsupported frequencies under magnification. We present ConeGaussian, a camera-model-agnostic anti-aliasing framework for Gaussian ray-based rendering. Instead of defining the pixel filter on a camera-specific image plane, ConeGaussian constructs an anisotropic footprint directly from neighboring rays produced by the camera's native inverse mapping. We derive a closed-form response under a locally linear, depth-local, moment-matched approximation of the finite pixel footprint, while the same geometry defines a per-Gaussian training-frequency floor. Notably, by construction, our filtering principle can be used unmodified across calibrated central camera models and multiple Gaussian ray-rendering backbones. Additionally, unlike in mip-splatting, our scene-space frequency floor and filtering enable trivial composition at render time, allowing us to remove excess blurring. On pinhole and strongly distorted fisheye captures, ConeGaussian consistently improves two distinct ray-based backbones, by up to 4.3 dB at 1/8 resolution, and reduces fisheye LPIPS by 30% where perspective screen-plane footprint formulations are not directly applicable.
Read the full abstract

All authors Deheng Zhang, Letian Shi, Runyi Yang, Zhendong Li, Lei Sun, Kanzhi Wu, Ajad Chhatkuli, Danda Pani Paudel, Luc Van Gool

Artificial Intelligence

Converge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian Optimisation

Multi-objective Bayesian optimisation (MOBO) is a sample-efficient approach for optimising expensive black-box functions with multiple objectives. In MOBO, the goal is to adequately approximate the Pareto front; that is, to obtain a high-quality solution set with 1) good convergence (closeness to the Pareto front) and 2) good diversity (spread across the Pareto front). Existing MOBO methods typically aim to accomplish these two tasks simultaneously, i.e., driving the search towards the Pareto front while maintaining a diverse set of nondominated solutions, such that the solutions, ideally, can gradually approach the entire front. When sufficient search budgets are available, this approach is effective. However, considering both convergence and diversity throughout the search is not easy and requires careful design. Under very tight budgets, there may not be enough solutions generated to be able to simultaneously approach the entire Pareto front. To address this issue, this paper proposes a converge-then-diversify (CTD) approach that decouples convergence and diversity into two stages. In the first stage, CTD focuses on convergence, aiming to quickly drive the search toward a single point on the Pareto front. In the second stage, CTD focuses on diversity, aiming to spread solutions across the front. We present two simple instantiations of CTD by using widely adopted acquisition functions in the area. Experimental results show that, across all 446 pairwise comparisons, CTD statistically outperforms state-of-the-art methods in 72.9% of the cases, performs equivalently in 21.1%, and is statistically worse in only 6.1%, with the advantage being particularly evident in settings with very tight evaluation budgets or in high-dimensional problems.
Read the full abstract

Machine Learning

Task-Aware Federated Fine-Tuning for MoE-based Large Language Models

Mixture-of-Experts (MoE) has become a widely adopted architecture for Large Language Models (LLMs), as it improves model capacity while limiting computational overhead through sparse expert activation. This property makes MoE-based LLMs particularly attractive for resource-constrained distributed environments. However, federated fine-tuning of MoE-based LLMs remains challenging under heterogeneous client data. Since clients often correspond to different task preferences, directly aggregating their local updates may weaken expert specialization and introduce conflicting update directions on shared experts. To address these challenges, we propose FedTAR, a task-aware federated fine-tuning method for MoE-based LLMs. FedTAR establishes the association between local updates and task preference via routing outputs. Specifically, we apply Singular Value Decomposition (SVD) to both routing features and local updates to extract low-dimensional task coordinates and update directions. Based on the task coordinates, FedTAR performs intra-cluster aggregation among clients with similar task preferences and inter-cluster aggregation across different task groups. The aggregated update is then reconstructed through the learned task-to-update mapping, ensuring that the final update remains aligned with task-specific optimization directions. In this way, FedTAR preserves expert specialization and mitigates destructive interference among heterogeneous clients. We evaluate FedTAR on four benchmark tasks under different non-IID settings. Experimental results demonstrate that FedTAR consistently outperforms strong federated fine-tuning baselines and achieves state-of-the-art performance.
Read the full abstract

All authors Tingqi Wang, Hongyu Ke, Haoxin Wang, Rafal Angryk, Zhipeng Cai