Probability
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Tuesday 8 September
In this work we establish a connection between the min-max optimal matching in the critical dimension $d=2$ and a simpler, geometrically linear, action of random curves in a Brownian potential. This is done by following Leighton and Shor and working with the dual formulation, which we approximate by a problem of isoperimetric-type with a random volume term of white-noise character. This allows us to extract the leading-order term in the asymptotics of the expected cost, sharpening previous results.
Representation Theory
The Auslander-Reiten conjecture for algebras with radical cube zero
Let $A$ be a split finite-dimensional algebra over a field whose radical $J$ satisfies $J^3=0$, and let $s$ be the number of isomorphism classes of simple $A$-modules. We prove that a non-projective module $M$ with ${\rm Ext}_A^i(M,A)=0$ for all $i>0$ has a non-zero self-extension in some degree between $1$ and $3s+1$. In particular, $A$ satisfies the Auslander–Reiten conjecture, which asserts that every self-orthogonal generator is projective. As a consequence, every finite-dimensional algebra over an algebraically closed field with radical cube zero satisfies the Auslander-Reiten conjecture.
Read the full abstract
Robotics
HiBRIDGE: A Hierarchical Bayesian Neural Network Framework for Interpretable Dialogue Management in Group-Robot Interaction
In multi-party human-robot interaction, a robot must continuously decide whom to address and what to say to participate effectively in the conversation. In real-world interactions, this is challenging because several behaviours may be plausible at the same time: a robot might continue a topic with one participant, involve another through a question, or address the whole group, with the appropriate choice depending on both whom it addresses and the interaction context. Current approaches remain limited in representing uncertainty when several behaviours are plausible and in structuring decisions into semantically meaningful intermediate steps that make robot decisions easier to interpret. Addressing these, we present HiBRIDGE, a hierarchical Bayesian neural network framework for group-robot dialogue management. Its Bayesian formulation enables uncertainty-aware prediction and robust learning from limited interaction data, while the hierarchical approach formulates behaviour selection as a structured, multi-stage decision process. We further use decision-tree surrogates to investigate whether this structure can support more interpretable explanations. Across three offline group-HRI datasets, our findings show that Bayesian formulations outperform their deterministic counterparts and several state-of-the-art baselines. Next, through an online study (N=20), we show that explanations derived from the hierarchical model are rated as more helpful for understanding robot behaviour and are preferred over those derived from the flat model. Finally, through our in-person study (N=12), we demonstrate the feasibility of HiBRIDGE for autonomous real-time group interaction, with both hierarchical and flat Bayesian variants positively perceived. Overall, HiBRIDGE combines strong predictive performance with a structured decision process that supports more interpretable explanations of robot behaviour.
Read the full abstract
Systems and Control
Entropic Risk-Sensitive Evolutionary Learning and Equilibrium Selection in Coordination Games
We study risk-sensitive evolutionary learning dynamics and their long-run equilibrium selection behaviors in coordination games. Agents' risk attitudes enter through the classical entropic risk measure, which evaluates opponent-induced payoff uncertainty and feeds into noisy best responses under two standard revision protocols: best response with mutations and logit choice. We first analyze $2\times 2$ coordination games in both single-population symmetric and two-population asymmetric settings. In the single-population setting, unlike the risk-neutral case where the dynamics are known to favor the risk-dominant equilibrium, we show that risk sensitivity can change the stochastically stable outcome: a greater risk-seeking attitude favors the payoff-dominant equilibrium, while a greater risk-averse attitude favors the maximin equilibrium. Thus, the population's risk attitude may act as a control knob for long-run equilibrium selection. In both population settings, we also identify a robust regime: any super-dominant equilibrium is stochastically stable for all risk attitudes, under both protocols, and across populations. We further extend the single-population analysis to symmetric $k$-action games, which include symmetric $k$-action coordination games as a special case, under risk-sensitive best response with mutations. In this setting, we show that, for sufficiently large populations, sufficiently risk-seeking agents uniquely select the strongly payoff-dominant equilibrium when it exists, whereas sufficiently risk-averse agents uniquely select the strongly maximin equilibrium when it exists. These results show that entropic risk sensitivity may serve as a systematic mechanism for steering equilibrium selection in evolutionary games, beyond the classical risk-neutral benchmark.
Read the full abstract
Physics and Society
HierX: Fast Multi-Scale Distance-Decay Interaction on Million-Node Networks
Many models across the sciences require global distance-decay interactions on large sparse networks. Gravity models, accessibility measures, spatial economic models, and network influence processes all evaluate distance-weighted potential fields: each location accumulates contributions from every other location, weighted by a decaying function of the shortest-path travel cost between them. Computed directly, such aggregate fields require the dense matrix of all pairwise network costs, which scales quadratically in time and memory. Common approximations either discard long-range contributions or fail when interaction is governed by network distances rather than geometric proximity. We introduce HierX, a hierarchical sparse-plus-correction operator for distance-decay potential fields on networks. HierX constructs multi-scale representative layers with explicit correction terms ensuring each location pair contributes exactly once at the finest available resolution. Under bounded-growth assumptions common in spatially embedded networks, applying HierX scales as O(n log n). Systematic benchmarks confirm quasi-linear scaling to 100,000 nodes. In head-to-head comparison with distance cutoff truncation and Nystrom low-rank approximation on 25,000-zone networks, HierX achieves 5-9% RMSE at a fraction of the computational work; Nystrom degrades severely on steep decay kernels. Case studies compute population-weighted accessibility on the 2.58-million-node Great Britain driving network and the 1.77-million-node London pedestrian network: a one-time hierarchy construction (~1 hour) yields a compact reusable operator that then evaluates each national-scale accessibility field in under one second (~700 ms for Great Britain, ~150 ms for London). Open-source code and worked examples are provided.
Read the full abstract
Mesoscale and Nanoscale Physics
Observation of linear magnetoresistance and planar Hall effect in the Dirac semimetal PdTe
PdTe is a Dirac semimetal that also exhibits superconductivity, providing an intriguing platform to explore topological superconductivity and unconventional magnetotransport phenomena. While the superconducting properties of PdTe have been extensively studied in recent years, the detailed magnetotransport phenomena have remained unexplored. Here, we present the first observation of linear magnetoresistance (LMR) and the planar Hall effect (PHE) in a high-quality single crystal of PdTe. We observe temperature-dependent unsaturated LMR in both in-plane ($B \parallel I$) and out-of-plane ($B \perp I$) configurations. The magnetoresistance (MR) shows a crossover from parabolic to linear dependence at the critical field Bc, and detailed analysis indicates that disorder-driven mobility fluctuations are the origin rather than the Abrikosov quantum-limit mechanism. Furthermore, prominent PHE has been observed by rotating the magnetic field within the plane of the sample. The detailed analysis of the field and temperature dependence of PHE-amplitude, along with the parametric plot, suggests that PHE originates predominantly from the anisotropic orbital magnetoresistance rather than the chiral anomaly. Our results demonstrate that although the Dirac point is close to the Fermi level, the observed LMR and PHE in the Dirac semimetal PdTe can be understood within a semiclassical transport framework, highlighting the importance of distinguishing between topological and conventional classical transport mechanisms in topological materials
Read the full abstract
Mesoscale and Nanoscale Physics
Wafer-Level 1/f Noise Characterization of Scaled P-Type Nanosheets and Comparison with Planar HKMG Transistors
We present a comprehensive wafer-level 1/f noise study of scaled Gate-All-Around (GAA) p-type nanosheet transistors and benchmark them against large-area planar High- ąppa} Metal Gate (HKMG) devices to assess the impact of device architecture on reliability. By statistically analyzing noise data from 188 nanosheets, we extract the effective trap density in the dielectric and compare it to that of planar transistors with identical gate stack, comparable processing thermal budgets, and no specific reliability anneals. The effective trap density extracted from 1/f noise is comparable in planar and nanosheet architectures. This indicates that transitioning to GAA nanosheets does not increase noise, which remains governed by gate stack quality.
Read the full abstract
Robotics
BIDETA: Brain-Inspired Data-Efficient Tactile Adaptation for Unseen Sensors
Vision-based tactile sensors provide high-resolution contact information for robotic perception and contact-rich manipulation, advancing embodied intelligence through more reliable physical interaction. However, device-specific sensing mechanisms cause tactile foundation models to degrade on unfamiliar hardware. Existing cross-sensor methods often require calibration data, paired observations, or iterative training. To address this problem, we propose Brain-Inspired Data-Efficient Tactile Adaptation (BIDETA), a gradient-free framework that uses a frozen tactile encoder and a few labeled target contacts to jointly predict labels for an unlabeled query batch. Inspired by the brain's rapid sensory adaptation, BIDETA combines rapid support memory, support-conditioned spectral graphs, and reliability-gated recurrence to preserve pretrained representations, repair sensor-dependent feature neighborhoods, and integrate reliable cross-query evidence. Experiments on SITR, TacVerse Shape, and TacQuad show that BIDETA substantially improves adaptation to unknown sensors: with only 10% labeled target data on SITR, it raises mean Sparsh accuracy from 6.86% for the frozen source classifier to 87.09%, exceeding the strongest implemented prior comparison by 47.22 percentage points, and these gains generalize across datasets, pretrained backbones, and tactile tasks. In the SITR timing benchmark with TVL, BIDETA also achieves approximately 20x faster target-sensor adaptation than the best baseline. BIDETA thus offers a gradient-free, data-efficient route to deploying tactile models on new hardware.
Read the full abstract
Sound
X2Streaming-ASR: wait when uncertain, emit when ready for streaming ASR
Streaming automatic speech recognition (ASR) for real-time voice agents and full-duplex dialogue must provide accurate partial transcripts with low commit latency. Existing systems commonly use a fixed chunk size, look-ahead, or target delay, or encourage emissions near estimated acoustic boundaries. These approaches do not directly optimize how much additional context to use at each output position under a single-pass, hard-commit constraint. We propose X2Streaming-ASR, which decomposes streaming recognition into when to commit and what to commit. Its three-stage training procedure first establishes streaming recognition ability, then warm-starts the commit policy with automatically probed trajectories, and finally refines the policy using character-level, segment-assigned group-relative rewards for recognition accuracy and latency. Across ten Chinese and English test sets, X2Streaming-ASR attains the lowest mean commit latency relative to forced-aligned endpoints, 32–109ms on Chinese characters and 12–85ms on English words, while recognition accuracy remains comparable to existing systems.Code is available at https://github.com/X-Square-Robot/X2Streaming-ASR.
Read the full abstract
Machine Learning (Statistics)
Optimal estimation for Functional Linear Regression with Noisy Discretized Data
In this paper, we consider the scalar-on-function linear regression model under a realistic sampling scheme in which the functional covariates are observed on a regular grid and contaminated by additive noise. We propose a two-step estimation procedure: first, the underlying curves are reconstructed from the discrete noisy observations using a Fourier-based projection method; second, the slope function is estimated by a penalized least-squares criterion over finite-dimensional trigonometric spaces, with data-driven selection of the model dimension. We establish oracle-type inequalities for the prediction error, both with respect to the reconstructed curves and to the true latent curves. Under regularity assumptions on the slope function and polynomial decay of the eigenvalues of the covariate, we derive convergence rates for the prediction error and show that our estimator attains the minimax rate when the number of grid points is sufficiently large. Finally, the proposed method is illustrated on simulated data and on a real meteorological dataset.
Read the full abstract
Combinatorics
Extremal Lin–Lu–Yau Curvature: Graph Density, Girth, and Short Cycles
We develop a variational framework for extremal Lin–Lu–Yau curvature under positive edge reweighting with the combinatorial metric fixed. For a connected locally finite graph $G$, define \[ \Kmax(G) := \sup_{w>0}\inf_{e\in E(G)}\kappa_{\LLY}^w(e). \] This invariant relates edgewise discrete curvature to global combinatorial and topological data. In the high-girth regime, we prove for every connected locally finite graph the exact identity \[ \Kmax(G)=\frac4{\mad(G)}-2, \] with the convention $1/\infty=0$. The formula is also equivalent to a normalized Euler-characteristic density: \[ \Kmax(G) = \inf_{\substack{H\subseteq G\text{ finite}\\E(H)\ne\varnothing}} \frac{2(c(H)-\beta(H))}{|E(H)|}. \] For arbitrary locally finite graphs, a nonnegative short-cycle surplus yields girth-dependent density bounds, with rigidity results in the finite case. We also classify the locally finite graphs of girth at least $6$ admitting a positive edge weight with nonnegative curvature on every edge: they are precisely the trees and the finite cycles.
Read the full abstract
Robotics
Learning to build covering structures with continuous adjustments
Robotic construction offers the potential to use materials more efficiently and create complex geometries, but current methods rely on rigid, high-precision plans that cannot accommodate the tolerances, inaccuracies, and unexpected changes inherent in physical fabrication. In this work, we introduce a reinforcement learning approach that forgoes predefined plans entirely, instead generating construction sequences adaptively as the structure is built. Our method operates on graph-structured state representations and a mixed (parameterized) action space, requiring both discrete block selection and continuous placement parameters. Because the stability simulation of a structure is computationally heavy, we develop an efficient exploration strategy by incorporating unilateral edges into graph neural networks, extending soft actor-critic (SAC) to this hybrid setting. We evaluate our algorithm, HSAC, against the prior method hybrid-PPO (HPPO), demonstrating significantly higher asymptotic performance and good sample efficiency. We also demonstrate HSAC's robustness to hyperparameter choices and its exploration capability, handling up to 10 discrete actions without performance degradation. Finally, we validate our approach on a physical two-robot setup, successfully building a spanning arch with 3D-printed blocks in closed-loop execution, confirming that policies trained in simulation transfer to real hardware.
Read the full abstract
High Energy Astrophysical Phenomena
Constraints on the Low-frequency Radio Emission of the Galactic FRB Source SGR 1935+2154
We present a search for radio pulses from the Galactic magnetar SGR 1935+2154, a well-known source of fast radio bursts (FRBs), at $\sim$110 MHz using the Large Phased Array (LPA) of the Pushchino Radio Astronomy Observatory. Data from two active periods in 2020 (March – May and September – November, with $\sim 3.5$ minutes of daily coverage) were analyzed with new methods tailored to both FRB-like single pulses and pulsar-like periodic signals. No significant FRB-like pulses were found. Using Monte Carlo simulations, $3\sigma$ upper limits were derived for the burst rate: for a log-normal energy distribution the limit is $\sim$${10}^{1.5}~{\rm{d}}^{-1}$ for a mean of average monochromatic isotropic luminosity $L_{\nu{\rm ,mean}}\sim1.3\times{10}^{29}~{\rm{erg~s^{-1}~ {Hz}^{-1}}}$ and a natural log-space scatter of $\sigma\sim0.85$; while for a power-law distribution it is $\sim$${10}^{1.8}~{\rm{d}}^{-1}$ for an index $\beta\lesssim3.0$ and a minimum average monochromatic isotropic luminosity $L_{\nu{\rm{,min}}}\lesssim0.7\times{10}^{25}~{\rm{erg~s^{-1}~{Hz}^{-1}}}$. When folded at the known 3.24781628 s period of SGR 1935+2154, a weak pulse was noted (S/N $<$ 3.16), but the significance is insufficient for a secure detection of the pulsar-like emission signal. A conservative upper limit on the average monochromatic isotropic luminosity of any possible periodic emission is $2.08\times{10}^{19}~{\rm{erg~s^{-1}~{Hz}^{-1}}}$. Our results offer meaningful low-frequency upper limits on the burst rate of SGR 1935+2154, and hint for very faint pulsar-like radiation at meter wavelengths.
Read the full abstract
General Relativity and Quantum Cosmology
Static spherically symmetric electroweak models with fermions and gravity
In this article we provide a systematic reduction to static spherical symmetry for (classical) gauge theoretic models with structure group SU(2) x U(1), coupled to gravity. In particular, we derive the most general static spherically symmetric ansatz for the Einstein-Yang- Mills-Higgs-Dirac-Yukawa equation system. This model is reminiscent of the lepton sector of the Standard Model coupled to gravity. The resulting total system of ordinary differential equations is rather large, so that finding global solutions is non-trivial. For this reason, an open-source interactive solver GUI written in C++ is provided together with the article.
Read the full abstract
Instrumentation and Detectors
Event-by-Event Space-Charge Gain Suppression in LGADs and Its Impact on Time Resolution
Low Gain Avalanche Detectors (LGADs) achieve time resolutions below 50 ps through internal charge multiplication, yet the gain is not a fixed device property as the space-charge of the multiplied electron-hole density screens the gain layer field, which suppresses further impact ionization. While TCAD simulations reproduce the average gain suppression, their deterministic treatment of a continuous charge density cannot capture the event-by-event fluctuations that determine the timing performance. We present a Monte Carlo study of space-charge-induced gain suppression in LGADs using the Garfield++ framework, in which the dynamic field of the drifting carriers, computed from a coarse-grained system of charged rings, is superimposed on TCAD-computed static field maps of Centro Nacional de Microelectrónica (CNM) and Hamamatsu Photonics (HPK) produced sensors while preserving single carrier avalanche statistics. The simulated gain reproduces published two-photon absorption-transient current technique measurements across two orders of magnitude in injected charge, including the dependence on the injection depth. For minimum ionizing particles, the event-by-event coupling between deposited charge and gain compresses the Landau tail of the multiplied charge, narrowing the signal arrival time distribution and improving the intrinsic time resolution of the HPK-S1 sensor investigated in this work from 53.9 $\pm$ 1.0 ps to 44.5 $\pm$ 0.9 ps. Treating the primary ionization as a stochastic sequence of charge clusters, it is further shown analytically that any gain suppression that decreases with the cluster size narrows the centroid time distribution.
Read the full abstract
Numerical Analysis
Off-Grid Point-Scatterer Localization from Sparse Limited-Aperture Data via Hankel Completion
We study the localization of $s$ distinct planar point scatterers from sparsely sampled single-incidence, single-frequency acoustic far-field data over a limited receiver aperture. Under the first Born approximation, the coherent data form a two-dimensional off-grid exponential sum and admit a low-rank Hankel representation. A limited aperture generally admits multiple aperture-admissible Hankel pencils with different lifted dimensions, sampling multiplicities, and Vandermonde conditioning. Hence, the choice of the Hankel pencil affects both data completion and subsequent localization. In this work, we introduce a unified framework for structured completion and off-grid localization based on general aperture admissible Hankel pencils. We establish exact recovery in the noiseless case and stable recovery under bounded noise from a number of random aperture samples that is linear in $s$, up to factors determined by the pencil geometry, source conditioning, and logarithmic terms, together with an explicit localization bound in terms of the completion error. The analysis separates the contribution of the pencil geometry from the conditioning induced by the source configuration and provides verifiable sufficient conditions for stable completion and MUSIC resolution. These results yield quantitative criteria for selecting sample efficient Hankel pencils under a prescribed physical aperture. Numerical experiments on connected and disconnected apertures demonstrate the effects of pencil geometry, sampling, and noise.
Read the full abstract
Analysis of PDEs
Long-term behaviour of the CAO-system: Optimal polynomial $H^1$-convergence to degenerate equilibria
This article is concerned with a definitive analysis of its long-term dynamics of the CAO-system. We establish the convergence of global strong solutions to constant equilibria in the full $H^1$-topology in the non-perturbative regime. Remarkably, we demonstrate that the CAO-system exhibits a polynomial rate of decay and we determine the optimal rate. This phenomenon stands in stark contrast to the exponential convergence typically expected for uniformly parabolic systems on bounded domains. This algebraic slowing is shown to be a purely nonlinear effect, driven by the emergence of a slow manifold (center manifold) induced by nonlinear coupling conditions. To the best of our knowledge, this provides the first instance of a global polynomial decay result for a non-gradient, non-local dissipative system on a bounded domain.
Read the full abstract
Machine Learning
MoEMB: Scaling Universal Multimodal Embeddings with Efficient Mixture-of-Experts Models
Universal multimodal embedding (UME) increasingly demands encoder's capacity for handling a broad range of tasks and modalities with increased complexity. Prior scaling methods either increase the representation size, retrieval effort, or scales the encoder into a heavy multimodal LLM. Recent works, such as Think-Then-Embed (TTE), explore scaling via reasoning tokens. However, embedding models are hard to scale up: increasing parameters directly tradeoffs for the large training batch size that contrastive learning needs, and retrieval has to be served under tight latency. Moreover, UME tasks are diverse in complexity, where scaling up embedders can bring significant redundant computation. In this work, we propose MOEMB, which instead scales UME along the expert axis through mixture-of-experts (MoE), growing encoder capacity while preserving single-vector, non-autoregressive encoding. Through a systematic study of the design space and training recipes for MoE-based UME, MoEMB sets a new state of the art on both MMEB-V2 and MRMR among models trained on public MMEB-family data: with only 3B active parameters, MoEMB surpasses TTE-based methods with >4x active parameters, using significantly less computes. To further improve the scalability and efficiency, we conduct the first comprehensive study of adaptive computation for MoE-based embedding, spanning diverse strategies across training-based and inference-only methods. Together, these results support expert scaling as an effective and efficient direction for UME, with adaptive computation further improving efficiency for MLLM-based embedding models towards large-scale retrieval and recommendation systems.
Read the full abstract
Information Theory
Linear programming bounds for binary and ternary LCD Codes
We derive linear programming (LP) bounds on the minimum distance of binary and ternary linear complementary dual (LCD) codes by imposing arithmetic constraints on their weight enumerators. Special values of the weight enumerator give finitely many Gauss phases, each of which yields linear equations in the ordinary weight-distribution variables. The resulting bounds strengthen the real-valued LCD counting LP without introducing additional variables; both the number of branches and the number of added equations per branch are bounded independently of the code length. Exact certificates establish strict improvements for 62 binary parameter pairs of length at most 20 and 39 ternary pairs of length at most 14. For four binary pairs, the bounds also improve the joint-weight-enumerator LP while using fewer variables per branch. These comparisons show that Gauss-phase information provides a compact strengthening of existing LP relaxations for LCD codes.
Read the full abstract
Signal Processing
Partial Superimposed Pilot-Aided Sparse Vector Transmission for High-Mobility URLLC
A partial superimposed pilot-aided sparse vector transmission (PSP-SVT) scheme is proposed for short-packet ultra-reliable and low-latency communications in high-mobility scenarios. Unlike conventional full superimposed pilot-aided SVT schemes, the proposed PSP-SVT scheme deploys only a few pilots over a subset of subcarriers. This sparse pilot structure is sufficient for basis expansion model based channel tracking while effectively reducing pilot-data interference. Based on the PSP pattern, an iterative receiver is developed to jointly perform channel estimation and data decoding. The reduced pilot interference in PSP-SVT provides more accurate initial channel estimation and data detection, thereby improving the subsequent iterative refinement and mitigating their error propagation. Moreover, the impacts of the number of PSPs and power allocation ratio on block error rate (BLER) performance are investigated to reveal the near-optimal pilot configuration. Simulation results show that the proposed PSP-SVT scheme outperforms existing full superimposed pilot-aided SVT schemes in terms of BLER with fast convergence speed.
Read the full abstract
Computer Vision
CoordFormer: Give Me Any Coordinates and I Will Give You Labels
Semantic segmentation on very-high-resolution images remains challenging due to the high computational cost and the difficulty of capturing fine-grained details. We propose CoordFormer, a novel coordinate-based architecture for semantic segmentation that predicts labels at arbitrary spatial locations through a Coordinate Decoder equipped with a Localized Cross-Attention mechanism. The decoder combines coordinate embeddings with high-resolution local patch features and interacts with global tokens extracted from a downsampled image processed by a ViT foundation encoder, enabling rich semantic context while preserving pixel-level precision. This design enables flexible inference at arbitrary resolutions while keeping memory low on very-high-resolution inputs, and supports an efficient semantic-edge-focused strategy that concentrates computation along boundaries, maintaining fine-grained accuracy while reducing latency and computational cost. CoordFormer achieves state-of-the-art performance on MaSS13K and outperforms comparably sized and higher-parameter methods on DIS5K and KPIs, demonstrating its effectiveness for high-quality, very-high-resolution semantic segmentation.
Read the full abstract
Computer Vision
TriCCOT: Tri-part Convolutional Conformal Transformer for Onboard Space Object Detection
Onboard object detection in Earth observation is constrained by limited computational resources and the absence of fully corrected imagery. While convolutional detectors are hardware-efficient, they often struggle to extract robust representations from raw and noisy data. Conversely, transformer-based models provide stronger global reasoning capabilities but remain difficult to deploy on FPGA accelerators due to quadratic attention complexity and non-compatible operations. We introduce TriCCOT, a tri-part architecture for robust and deployable onboard object detection. TriCCOT combines a convolutional region proposal network, a conformal prediction stage, and Aper-GATES, our hardware-friendly attention-based classifier. The region proposal network generates candidate bounding boxes, which are subsequently enlarged via conformal prediction, providing a distribution-free probabilistic coverage guarantee. The resulting crops are processed by Aper-GATES, which reformulates self-attention through convolutional projections, global channel statistics, and hardware-friendly gating operations, avoiding standard transformer operations that are poorly suited to CNN-oriented accelerators. Experiments on the DIOR and VDVRaw datasets demonstrate competitive detection performance and improved robustness to spatial blur and signal-dependent noise when compared to FPGA-compatible architectures. Finally, we report full deployment on a Xilinx Versal VCK190 FPGA without modifying the underlying DPU architecture, enabling unified CNN-Transformer inference for spaceborne embedded applications.
Read the full abstract
Combinatorics
Cayley Tournaments Simultaneously Critical for the Clique and Dichromatic Numbers
For a tournament $T$, let $\omega(T)$ be the minimum clique number among the backedge graphs of $T$, and let $\chi(T)$ be its dichromatic number. We give a template-lifting construction. It turns a $k$-template into a regular, vertex-transitive Cayley tournament that is simultaneously $(k+1)$-$\omega$-critical and $(k+1)$-$\chi$-critical. The output is also a $(k+1)$-template. Iterating the construction, we prove that for every $k\geq3$, there is a positive even integer $m_k$ with the following property. Every $N>1$ with $N\equiv1\pmod{m_k}$ is the order of a regular, vertex-transitive Cayley tournament that is simultaneously $k$-$\omega$-critical and $k$-$\chi$-critical. This proves a conjecture of Aboulker, Aubian, Charbit, and Lopes and gives a negative answer to their bounded-certificate question when the hypothesis is $\omega(T)\geq k$. We also find the clique number of a cyclic substitution when each block satisfies $\omega=\chi$. We then describe exactly when this substitution is $\omega$-critical if the blocks are $\chi$-critical and satisfy $\omega=\chi$.
Read the full abstract
Computer Vision
Charts Are Beyond Pixels: Probing for Layer-Wise Chart Understanding and Editing
Charts are structured visual compositions whose elements have distinct functional roles, semantic correspondences, and visibility relations. This structural view motivates evaluating whether models can understand and manipulate charts at the layer level. Existing chart benchmarks, however, primarily assess the correctness or fidelity of final outputs and do not directly evaluate these layer-wise behaviors. We present LayerWiseBench, a benchmark organized around three core concepts, layer attribution, layer binding, and visibility ordering, that structure its chart-understanding and chart-editing evaluations. Generated from executable chart programs, LayerWiseBench pairs each rendered chart with spatially aligned per-layer RGBA assets and construction-derived labels for functional roles, semantic bindings, and visibility relations. From this layer-wise representation, we derive controlled understanding questions, editing targets, reference images, and evaluation regions. It contains 2,800 source charts across 14 chart paradigms, from which we derive 7,329 layer-wise understanding questions and 53,791 instruction-guided editing variants. Among the evaluated VLMs, Qwen3.5-27B, which achieves the highest QA macro-average, obtains 93.04% accuracy on layer attribution and 97.46% on layer binding, but only 61.46% on visibility ordering. Across the four evaluated image editors, overall mIoU ranges from 1.49% to 4.93%, and visibility-constrained edits have the lowest mIoU for every editor, ranging from 0.37% to 2.00%. Taken together, these results identify tasks involving front-to-back relations between overlapping components as a recurring challenge across understanding and editing, motivating more explicit modeling of component identity and visibility relations.
Read the full abstract
Optimization and Control
A Lower Bound for the Heavy-Ball Method on Smooth Convex Functions
Can the classical Heavy-Ball method, with arbitrary horizon-dependent parameters chosen in advance, achieve Nesterov's $O(T^{-2})$ last-iterate rate on every smooth convex objective? We provide a negative answer. For every horizon $T\ge2$ and every predetermined schedule with nonnegative step sizes and momenta in $[0,1)$, there exists a convex $1$-smooth objective, with initialization distance at most one and zero initial velocity, for which the last iterate of the Heavy-Ball method satisfies \[ f(x_T)-f^\star=\Omega\!\left(\frac{1}{T^\alpha\log T}\right), \qquad \alpha=\frac{1+\sqrt5}{2}. \] Thus even fully nonstationary, horizon-dependent tuning cannot give the classical Heavy-Ball method a Nesterov-rate guarantee on the smooth convex class.
Read the full abstract
Quantum Gases
Exciton-polariton condensates in epsilon-near-zero cavities
We propose an epsilon-near-zero (ENZ) artificial cavity that can enhance light-matter coupling around the ENZ frequency and theoretically show the possibility of exciton-polariton condensation in the suggested platform. By using a rate equation model based on the driven-dissipative reservoir approach, we calculate the pump threshold for condensation and then compare the suggested platform's advantages/disadvantages with distributed Bragg reflector cavities. We also discuss the validity of the mean field theory within our model.
Read the full abstract
Astrophysics of Galaxies
The environmental dependence of the circumgalactic medium in a high-resolution cosmological simulation
There is increasing evidence from observations that the circumgalactic medium (CGM) of galaxies depends on the large-scale structure in which they are embedded. When probing the CGM in absorption using quasar sightlines, studies find an enhanced sky coverage in the CGM of galaxies in overdensities compared to galaxies in isolation. However, the exact reason for this environmental dependence is still unclear. In this work we aim to model for the first time the influence of the large-scale structure on the cool and warm ($T\sim10^{4-5}$ K) gas phases of the CGM. We use a high-resolution ($m_{gas}\approx 4.5\times 10^4$ M$_{\odot}$, $m_{dm}\approx 2.4\times 10^5$ M$_{\odot}$) cosmological simulation based on the EAGLE model of galaxy formation. We select all galaxies at $z=0$ with stellar mass $M_*>10^8$ M$_\odot$ and split them into galaxies in overdensities (group galaxies) and galaxies in isolation using a Friends-of-Friends algorithm. For these two samples, we investigate how the large-scale structure influences the physical properties of the CGM and the measured covering fractions of the cool and warm gas phases. When the two samples of group and isolated galaxies are matched in stellar mass, halo mass, and we use only central galaxies, we do not find any significant difference in the physical properties of the CGM and the measured covering fractions. However, when satellite galaxies are included, we recover the observational trends in the difference of covering fractions with the environment. The difficulty of recovering the observational trends shows the complexity of capturing the multiphase CGM in simulations. However, since our results concerning the admixture of satellites are independent of the employed subgrid physics, this work shows that central galaxies and satellites need to be disentangled in observational studies to clearly discern the role of the environment on the CGM.
Read the full abstract
General Relativity and Quantum Cosmology
Persistence of measurement-induced nonlocality in uniformly accelerating Unruh-DeWitt detectors
Uniform acceleration induces a thermal response to the Minkowski vacuum and can thereby modify quantum correlations. For measurement-induced nonlocality (MIN), a form of nonclassical correlation distinct from entanglement, previous field-mode analyses of bosonic fields found that it decreases with increasing acceleration and vanishes in the infinite-acceleration limit. This raises the question of whether the disappearance of MIN is a generic consequence of the Unruh effect or depends on the physical description of the accelerated quantum system. To address this question, we study two uniformly accelerating Unruh-DeWitt detectors interacting with a massless scalar field. We find that the response of MIN to the Unruh temperature depends sensitively on the initial detector state: it can decrease monotonically, vanish at an intermediate temperature and subsequently recover, or increase monotonically. Moreover, for a broad class of initial states, MIN approaches a nonzero value in the high-temperature limit. These results show that the suppression of MIN under strong acceleration is not universal. Instead, Unruh-induced detector dynamics gives rise to a state-dependent response in which measurement-induced nonlocality can be suppressed, restored, or enhanced.
Read the full abstract
Group Theory
On the unitary cohomology of semisimple groups
We study the continuous homology and cohomology of semisimple Lie groups with coefficients in arbitrary unitary representations. The case of irreducible representations was determined by Vogan and Zuckerman; we focus on reducible representations. The (co)homology splits into its Hausdorff and torsion parts. The Hausdorff part is governed by containment of irreducible cohomological representations. We show that the torsion part is governed by weak containment of irreducible cohomological representations. Precisely, we show that a unitary representation admits non-zero torsion if and only if there is a cohomological point which is not isolated in its support. We discuss in length examples of rank-$1$ groups, completely determining unitary cohomology for the group $\mathrm{SO}^\circ(n,1)$. For a simple Lie group, the first and fourth named authors showed that the first degree in which it obtains non-trivial cohomology for some unitary representation with no invariant vectors is related to the rank of the group. We discuss the analogous question regarding torsion cohomology, and show that the corresponding first degree could be much higher: in the presence of property (T), it is bounded below by the square root of the dimension of the symmetric space. A technical device that we use is the restriction to well chosen dense subgroups which satisfy finiteness properties, which we call cohomological witnesses. We combine it with results on the unitary cohomology of groups which satisfy finiteness properties. These results are of an independent interest.
Read the full abstract
Fluid Dynamics
Closed-Form Solution for Oscillatory Flow and Wall Shear Stress in Axisymmetric Corrugated Tubes
Pulsatile flow in corrugated tubes arises in hemodynamics and microfluidics, where oscillatory forcing and geometric constrictions jointly determine transport and wall loading. We extend Womersley's classical solution for oscillatory flow in a rigid circular tube to rigid axisymmetric tubes of slowly varying radius. Within the lubrication approximation, we derive closed-form expressions for the axial velocity profile, volumetric flow rate, phase lag, and wall shear stress at an arbitrary Womersley number. Three-dimensional lattice Boltzmann simulations are used to assess the regime of validity of the theory. The analysis shows that the local velocity profile varies strongly along the tube, ranging from plug-like in wide sections to more parabolic near bottlenecks. The cycle-maximum flow rate decreases with increasing corrugation, but this reduction weakens as pulsatility increases, reflecting a crossover from the quasi-steady scaling $\langle R^{-4}\rangle^{-1}$ to the high-frequency scaling $\langle R^{-2}\rangle^{-1}$. The wall shear stress is maximal at the bottleneck and decreases with Womersley number. For sinusoidal corrugations, the bottleneck wall shear stress depends non-monotonically on the corrugation because of the competition between local shear amplification and global hydraulic resistance. Closed-form expressions for the time-averaged wall shear stress and oscillatory shear index further connect the theory to standard hemodynamic metrics.