Astrophysics of Galaxies
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Wednesday 2 September
Low surface brightness galaxies (LSBGs) are very faint objects in the universe whose formation and evolution have long remained an unresolved problem. We selected a late-type, face-on, and disk-dominated galaxy sample from the MaNGA survey, composed of 38 LSBGs and 216 high surface brightness galaxies (HSBGs) by adopting a central surface brightness threshold $\mu_{0}(g) = {\rm 22.0\ mag\ arcsec^{-2}}$. Based on maps derived from pipe3D products of star-forming areas in galaxies, we investigated the global and resolved star formation main sequence (SFMS and rSFMS) and radial profiles of the stellar mass density ($\mathrm{\Sigma_{*}}$), star formation rate density ($\mathrm{\Sigma_{SFR}}$), and specific star formation rate ($\mathrm{\Sigma_{sSFR}}$). The radial profiles of $\mathrm{\Sigma_{*}}$ and $\mathrm{\Sigma_{SFR}}$ present negative gradients, while those of $\mathrm{\Sigma_{sSFR}}$ present positive gradients. Compared to HSBGs, the LSBGs show similar, flatter, and steeper slopes in $\mathrm{\Sigma_{*}}$, $\mathrm{\Sigma_{SFR}}$, and $\mathrm{\Sigma_{sSFR}}$ gradients, respectively. The very flat $\mathrm{\Sigma_{SFR}}$ gradients of LSBGs, whose slopes range from -0.1 to -0.2, indicates uniform star formation throughout the whole galaxy, and the positive $\mathrm{\Sigma_{sSFR}}$ gradients reflects their “inside-out” growth. Our LSBGs and HSBGs follow the same global and resolved SFMS as the full sample, and $\mu_{0}$ shows no influence on both SFMS and rSFMS. For our full sample and two subsamples (HSBGs and LSBGs), the slopes of global SFMS and rSFMS are consistent, which indicates that the SFMS still hold on a kiloparsec scale and the star formation activities are regulated by local events. The treatment of different excitation sources in galaxies affects the slope of the SFMS and rSFMS, and using star-forming areas only is more advisable.
Computer Vision
ProSR: Semantic-Prototype-Guided Discrete Modeling for Physically Consistent SAR Super-Resolution
High-resolution Synthetic Aperture Radar (SAR) imagery is critical for precision analysis such as automatic target recognition, yet its acquisition is costly. Although generative image super-resolution (ISR) models offer a promising alternative, current smooth-approximation based diffusion frameworks often struggle to preserve the coherent scattering statistics, causing stochastic structural distortions that are less consistent with real SAR physics. To address this, we propose Semantic Prototype-Guided Super-Resolution (ProSR), reformulating SAR ISR as a semantically-guided discrete token prediction task within a quantized latent space. By mapping signal features to discrete scattering primitives, ProSR preserves the impulsive nature of SAR without over-smoothing. Furthermore, we integrate a Self-Supervised Learning backbone into SAR ISR to extract label-free semantic priors, overcoming label scarcity. Guided by these priors, we introduce Semantic-Aligned Detail Encoding to decouple high-frequency signals into discrete scattering primitives. In parallel, the Semantic Prototype Map Generator explicitly constructs semantic prototype maps, allowing Prototype-Map-Guided Attention to route the information flows within identical categories and mitigate inter-class interference. To validate our approach, we present a large-scale 0.25m resolution benchmark from the Umbra Open Dataset. Experimental results show ProSR achieves superior visual quality while preserving essential scattering characteristics required for practical SAR applications.
Read the full abstract
Sound
Removing Speech, Keeping Activities: Speech Removal for Privacy-Preserving Acoustic Sensing in Assisted Living
Acoustic sensing can monitor the daily activities of older adults without cameras or wearables, yet the possibility of capturing private conversations remains a barrier to its deployment. The engineering application addressed in this paper is privacy-preserving acoustic monitoring of daily activities in assisted living. Our contribution in artificial intelligence is a speech-removal front end: a U-Net that masks the short-time Fourier transform of ambient audio and returns a waveform in which speech is suppressed and the sounds needed for activity recognition are kept. The model is trained only on synthetic mixtures of public sounds and speech, so no speech has to be recorded in the target homes. We evaluate it with group-disjoint cross-validation on environmental and domestic sound datasets mixed with reverberant speech, on speech from unseen corpora and on real home recordings, and measure privacy with three voice activity detectors, Whisper speech recognition and a speaker-verification attack. With speech 5 dB above the background, the macro-averaged F1 score of activity recognition is 0.574 and 0.661 after removal on the two datasets, compared with 0.233 and 0.178 without processing, and higher than that of speech enhancement, speech separation and detector-gated muting at every speech level. Whisper recovers 5.2% of the spoken words, compared with 89.0% without processing and a chance floor of 2.2%. On the real recordings, the speech proportion estimated by a voice activity detector falls from 8.5% to 2.7%. The method reduces, but does not eliminate, spoken content and speaker identity.
Read the full abstract
Cell Behavior
3D hybrid cellular Potts model with a discrete deformable fiber network: modeling cell contraction and extracellular matrix remodeling
The extracellular matrix (ECM) is a fibrous and dynamic network that plays a critical role in development, homeostasis, and disease. Cells both respond to and remodel the ECM, engaging in a mechanical reciprocity that shapes tissues. To study these interactions, computational models have been developed that simulate either ECM mechanics or cell behavior. The Cellular Potts Model (CPM) is a flexible cell-based framework that has been extended to many biological processes, including coupling to a discrete deformable fiber network. So far, however, this extension has only been applied in two dimensions, even though a three-dimensional setting is biologically more relevant. Here, we present a 3D hybrid framework that couples the CPM with a discrete and deformable representation of the ECM. This model enables explicit simulation of cell-induced ECM remodeling, including fiber reorientation and matrix densification. By incorporating contractile forces through static adhesion points, the model captures how ECM elasticity and fiber stiffness influences cell shape. The simulations show that ECM stiffness, controlled by fiber crosslinking, resists contraction and reaches equilibrium, while crosslink density modulates fiber alignment and local matrix accumulation. This framework provides a versatile platform for studying cell-ECM mechanics and supports future studies of multicellular behavior in realistic 3D environments.
Read the full abstract
Optics
Static and breathing dynamics of optical bullets in nonlinear GRIN fibers
We explore the properties of spatiotemporal solitons within nonlinear graded-index (GRIN) optical fibers, specifically addressing their formation, stability, and breathing dynamics. We solve the (3+1)-dimensional nonlinear Schrödinger equation employing a semi-analytical variational approach, which leads to a set of reduced coupled ordinary differential equations describing the bistable bullet dynamics. The stability of these states is assessed using the Vakhitov-Kolokolov criterion and linear stability analysis within the Kantorovich optimization framework. The breathing dynamics of the optical bullets are investigated adopting the potential formalism, revealing an oscillatory behavior and self-imaging dynamics in GRIN fibers. The analytical results demonstrate strong agreement with full numerical simulations, validating the robustness of our theoretical model. Our research enhances the understanding of stabilizing self-repeated 3D optical structures, which is beneficial for the development of advanced photonic technologies.
Read the full abstract
Machine Learning
Percolation Dynamics in Optimization: Variance Cascades and Nested Symmetry
We study the dynamics of Stochastic Gradient Descent (SGD), which is known to steer deep neural networks toward invariant sets that correspond to simpler subnetworks. How this steering unfolds over time remains poorly understood. We answer this by modeling the stochastic gradient flow (SGF) as a percolation process, in which nested architectural symmetries force subnetworks to merge in discrete blocks rather than by single-edge attachment. These structural transitions register as variance spikes in a macroscopic order parameter echoing physical phase transitions. We further state sufficient conditions under which the trapping argument carries over to Adam and AdamW under heavy-tailed gradient noise and measure them on a trained Transformer.
Read the full abstract
Quantum Physics
Implementation of quantum gates by Floquet analysis of kicked quantum system
Precise control of multi-qubit architectures remains a critical bottleneck in superconducting quantum processors. In this work, we investigate the synthesis of high-fidelity quantum operations and state transfer protocols within an extended superconducting linear chain, scaling from three to seven sites. Using Floquet theory, we model the periodic drive as a train of delta-like pulses, mapping the quantum control problem onto quasi-energy resonance conditions. Combining the Baker-Campbell-Hausdorff expansion with Floquet spectral decomposition, we analytically identify optimal driving parameters, refined via the Covariance Matrix Adaptation Evolution Strategy (CMA-ES). In the three-qubit architecture, this enables high-fidelity synthesis of the iSWAP gate. Extending to a seven-site chain, we implement periodic trains of finite-width Gaussian pulses to activate distinct double-excitation transport channels with ultra-short gate durations t_gate (~170 ns). This achieves a clear scale separation from energy-relaxation times (T1) typical of fixed-frequency transmon devices with tunable couplers, such as IBM Quantum hardware. Finally, we benchmark stability under realistic imperfections, revealing a heightened sensitivity to static parameter disorder at the sub-percent level (eta ~ 10^-3) driven by spectral crowding, and discuss how closed-loop topologies could mitigate this constraint. This framework bridges time-periodic control theory and practical quantum gate engineering.
Read the full abstract
Artificial Intelligence
Diagnosing with Insights: Structured Analysis of Agent Failures via Behavioral Abstractions
With the proliferation of LLM agents, the ability to understand and diagnose failures in agents is essential to achieving superior effectiveness and trustworthiness. As agent failures often manifest via long and complex trajectories, manually finding the needles in the haystack is untenable. However, traditional diagnosis techniques for software bugs can hardly address LLM agent failures, while completely relying on LLMs as the judge yields unreliable diagnosis results. To overcome these challenges, this paper presents AGENTSCOPE, a new neuro-symbolic approach for agent failure mode diagnosis. The key principle of AGENTSCOPE is to abstract agent behavior, based on its trajectories, into structured representations. Furthermore, AGENTSCOPE introduces the concept of neural invariants to specify agent behavior properties. AGENTSCOPE leverages LLM-guided reasoning atop the structured representation against neural invariants to pinpoint both the failure step and its type in the trajectory. We show the effectiveness of AGENTSCOPE on publicly available agent failure datasets (Who&When) and a more comprehensive dataset created by us (AgentErrata), where AGENTSCOPE significantly outperforms the current state of the art in fault localization and attribution accuracy. Our work shows that integrating structured abstractions with LLM-guided reasoning enables effective, reliable, and interpretable diagnosis for agent failures.
Read the full abstract
High Energy Physics (Theory)
Surprises of Klein paradox
We consider second quantized $1+1$ dimensional Dirac fermions with mass $m$ in background electric fields. This theory mimics the following situations. We consider two parallel capacitors with opposite directions of the electric field $E$ in them. The potential energy difference in one of them is equal to $V_0 > 2m$ and in the other to $V+V_0$, with $0<V<2m$. According to general physical understanding of the Schwinger effect, each capacitor separately will create pairs and there will be oppositely directed currents, which will not be equal to each other. The total current will not vanish. In this paper we show that this is too naive an observation: the limiting expectation value of the current operator, as $t\to +\infty$, is actually zero, if the above form of the potential with $0<V<2m$ is instantaneously turned on. The non-vanishing current is obtained only if $V>2m$. We also consider the same theory in the presence of the infinite wall and a deep well $V_0>2m$. We show that in such a potential the limiting expectation value of the current is zero in the Gaussian approximation. We discuss the relation of these observations to the Schwinger effet and to the decay of supercritical nuclei.
Read the full abstract
Computer Vision
Information Density Imbalance in Visual Object Detection
In object detection, the number of instances is typically used to determine whether a dataset exhibits a long-tailed distribution, implicitly assuming that the model will perform poorly on categories with fewer instances. This assumption has led to extensive research on category bias in datasets with imbalanced instance numbers. However, even in datasets where instance numbers are relatively balanced, models still exhibit category bias, indicating that instance count alone cannot explain this phenomenon. In this work, we first introduce the concept and measurement of information density. We then observe a significant negative correlation between a category's information density and its accuracy, and we investigate how the training process impacts this relationship. Empirical studies suggest that information density imbalance may be a potential source of category bias. To preliminarily validate the potential of information density, we made simple improvements to three advanced object detection loss functions using this concept. Experiments on the Pascal VOC, COCO-LT, and LVIS datasets demonstrate that information density can significantly reduce model bias while effectively enhancing the overall performance of existing loss functions. This study provides a new perspective for understanding the generalized bias phenomenon in object detection models and offers new tools for designing fairer loss functions and training strategies.
Read the full abstract
Materials Science
Magneto-Structural Coupling Enables Cryogenic Cation Redistribution in a Spinel Oxide
Ionic transport in oxides is generally frozen at cryogenic temperatures, where thermal energy lies far below typical cation-migration barriers. Neutron powder diffraction reveals progressive Fe/Mg redistribution between tetrahedral (A) and octahedral (B) sites in the spinel Mg0.5Fe0.5TiFeO4 upon cooling from 200 K to 5 K. A-site Fe occupancy increases toward near completion at 5 K within Rietveld resolution, while Ti remains on the B site. This exchange coincides with complex magnetic correlations rather than a classical thermally activated window. Low-temperature magnetostrictive volume changes indicate strong spin-lattice coupling, but do not identify magnetostriction as the sole thermodynamic driver. Room-temperature high-pressure X-ray diffraction produces the opposite occupancy trend, showing that volume contraction alone cannot explain the cryogenic site exchange. These results point to magneto-structural free-energy minimization as a plausible mechanism for unlocking cryogenic cation mobility in a correlated spinel oxide.
Read the full abstract
Multimedia
The Missing Temporal Link: Temporal Context Routing for Script-Driven Audio-Video Generation
Joint audio-video generation models have made substantial progress in visual quality and audio-visual synchronization. However, they still provide limited control over when shot transitions occur and dialogue is spoken. This limitation constrains their application in script-driven content creation, where timing errors can undermine narrative coherence and the viewing experience. Current joint generators align video and audio representations on a shared temporal axis, yet the precise timing of shots and dialogue specified in a structured prompt is encoded only in the prompt's text representation and remains unaligned with the temporal coordinates of either modality. Consequently, video and audio may remain synchronized with each other while both fail to follow the script timeline. This mismatch motivates us to extend temporal alignment beyond video and audio to include the structured script. We therefore introduce Temporal Context Routing (TCR), which maps the script timing onto the shared temporal axis of video and audio generation and routes each prompt's guidance to the corresponding positions in both modalities. Compared with the baseline on 200 test scripts, TCR reduces Shot Boundary MAE by 96%, from 1.11 s to 0.042 s, and raises Dialogue Acc@0.5 s from 28.3% to 84.1%. TCR achieves these improvements while maintaining visual quality and audio-visual synchronization comparable to those of the baselines. A user study further shows that participants prefer TCR on all five evaluated dimensions.
Read the full abstract
Computation and Language
NE-R1: Enhancing Named Entity Recognition Model via Reinforcement Learning
Named Entity Recognition (NER) has achieved substantial progress since the advent of large language models (LLMs). Nevertheless, the recognition of long-tail and domain-specific entities remains challenging due to the deficiency in parametric knowledge. Retrieval-augmented generation (RAG) offers a promising remedy by injecting external knowledge, but it also introduces noise and unnecessary cost when dealing with familiar cases. In this paper, we propose NE-R1, a novel framework for adaptive retrieval-augmented NER. We design a "retrieval-on-demand" mechanism for NER. Then we integrate it into models by a two-stage training method: (1) multi-task instruction tuning initialization; (2) end-to-end RL optimization with CoT. To achieve reasonable selection between parameterized and external knowledge, we design a multi-dimensional reward considering both accuracy and retrieval benefit. NE-R1 achieves state-of-the-art performance on various benchmarks, with an average F1 score gain of 2.52% in in-domain evaluation and 1.18% in zero-shot cross-domain evaluation.
Read the full abstract
Populations and Evolution
Beyond species area curves: a theoretical approach to the relationship between diversity and area
Species area curves, which describe the number of species present as a function of area, have long been used to understand biodiversity and inform conservation efforts. While understanding the number of species is important, it leaves out information about the population sizes of each species. In this work, we examine the relationship between the effective number of species from different diversity indices, like Simpson's index and the Shannon index, and area. These effective number of species are also referred to as Hill numbers. Using a spatial and neutral model that has previously been used to understand species area curves, we show through a combination of theory and simulations that the relationship between the effective number of species and area is linear when the area is sufficiently large and resembles a power law for smaller areas.
Read the full abstract
Human-Computer Interaction
Towards a Foundational Ontology for Identifying and Resolving Contradictions in Dialogue-based Human-Robot Interactions
Existing Human-Robot Interaction (HRI) literature has focused on identifying and structuring errors, failures, conflicts, and knowledge issues (called in this work as contradictions) in domain-specific dialogue-based interactions. However, there is still lack of a formal computational framework to represent and define these contradictions, interoperable and usable across HRI and human-agent interaction (HAI) domains. Thus, this research project aims to capture, represent, and evaluate the notion of (1) dialogue-based collaborative interaction and (2) related contradictions in a foundational ontology. METHONTOLOGY, a systematic approach to build domain-independent ontologies was applied. In the conceptualisation stage of the presented ontology, concepts and models from Activity Theory were used. Preliminary results presented in this short article are: (i) Natural language definitions of dialogues and related contradictions in HRI, (ii) Set Theoretic definitions of dialogues and contradictions, and (iii) First Order Logic (FoL) formulation of the contradiction concepts and three novel principles guiding dialogue-based interactions between humans and robots. In summary, we report on ongoing work to develop a foundational ontology based on Activity Theory called Activity Theory-based foundational ontology (ATFOt) to capture and represent the notion of contradictions in HRI.
Read the full abstract
Optimization and Control
On the invariance of risk-sensitive LQR gain under input randomization
This paper shows that the optimal gain of the risk-sensitive linear quadratic regulator (LQR) problem is invariant under input randomization when the cost penalizes the realized noisy input. This may appear counterintuitive at first glance because certainty equivalence does not hold for risk-sensitive LQR and input randomization inflates the effective process noise. Nonetheless, the gain is preserved because the input noise enters not only the system dynamics but also the cost functional, and its total effect eventually vanishes. Consequently, the optimal gain and the Riccati recursion need not be recomputed, and the increment in the optimal cost can be readily evaluated in closed form. This result facilitates the use of risk-sensitive LQR in applications that employ input randomization for privacy or exploration, such as dynamic watermarking for replay attack detection.
Read the full abstract
Probability
Farthest-cell triplet entropy: high-dimensional shell limits and hyperbolic curvature amplification
We introduce farthest-cell triplet entropy, the conditional Shannon entropy of the farthest-prototype label given three random prototypes. For independent queries and prototypes, its estimator records only the farthest label, not coordinates or numerical distances. The statistic is bounded by $\log 3$, is invariant under common strictly increasing transformations of the dissimilarities, and has an exact mutual-information interpretation. In high-dimensional isotropic radial models $X_d=R_dU_d$, the Euclidean ordering reduces to scores $\lambda_d\xi_{i,d}-Z_i$, where $\lambda_d=\sqrt d\,\operatorname{sd}(R_d)/\mathbb{E} R_d$ and the $Z_i$ are independent standard Gaussian variables. This gives angular-dominated, intermediate, and radial-dominated entropy limits $\log 3$, $H_{\infty}(\lambda;F)$, and $0$. In hyperbolic space of curvature $-\kappa_d^2$, the same master curve appears at $\lambda_{d,\mathbb H}=\sqrt d\,\tau_d A(s_d)$, where $\tau_d=\operatorname{sd}(R_d)/\mathbb{E} R_d$, $s_d=\kappa_d\mathbb{E} R_d$, and $A(s)=s\coth s$. With a calibrated radial law and $\tau_d$, and a monotone operating interval, entropy inversion identifies the scale-invariant target $s_d^2$; absolute curvature requires an external length unit. CPU simulations give Euclidean and hyperbolic master-curve RMSEs of $0.0164$ and $0.0209$. Inversion from observed synthetic latent coordinates has a median relative error in $\kappa$ of $6.6\%$, while angular anisotropy increases this error to $68.9\%$. Thus the entropy statistic is comparison-based, whereas curvature recovery remains model-calibrated, is not graph-only, and is not robust to anisotropy.
Read the full abstract
Quantum Physics
DMRG using Belief Propagation
Tensor networks have attracted much attention as a powerful tool for modeling quantum many-body systems. Their contraction is a significant challenge, however, especially in highly connected networks, as memory requirements become prohibitive and the optimal contraction order is increasingly hard to find. The belief propagation (BP) algorithm has emerged as an alternative to exact contraction. Being formulated in a graph-agnostic way, it offers great flexibility, but its accuracy suffers in the presence of loops. In this work, we combine BP with the DMRG algorithm to solve ground-state problems, thereby extending DMRG to higher dimensions and arbitrary lattices. We demonstrate the viability of BP-DMRG on the transverse-field Ising model on a $2\times 2$ hexagonal lattice, finding that it produces states with a fidelity between $0.9$ and $0.99$ to the true ground state, and energy estimates with a relative error between $10^{-2}$ and $10^{-3}$. Additionally, BP-DMRG can find ground states on randomly generated lattices, with fidelity improving as the transverse field increases. We conclude with a discussion of the limitations we encounter when using belief propagation, highlighting that the TFI tensor network operators lead to larger errors during BP iterations in BP-DMRG.
Read the full abstract
Soft Condensed Matter
Collective Cell Fluidity Controls Active Prestress Transmission in Cell-Extracellular-Matrix Tissues
Tissues are active composites in which multicellular collectives and extracellular matrices mechanically reorganize one another. We develop a three-dimensional micromechanical model that couples deformable, rearranging cell clusters to a disordered network of semiflexible fibers through a dynamic, force-generating interface. Cell clusters are represented as solid-like or fluid-like vertex-model spheroids and coupled to the matrix by passive or contractile linkers renewed as the cluster boundary reorganizes. Matched intact, voided, and passive-linker controls separate cavity formation, interfacial tethering, and active loading. At small strain, passive tethering provides modest reinforcement, whereas active contraction prestresses and strongly stiffens the matrix. Solid-like clusters preserve coherent force transmission and exhibit an excess modulus scaling approximately as $|\sigma|^{1.4}$ across changes in activity, cluster size, and cluster number. Fluid-like clusters undergo greater interfacial renewal, producing weaker and nonmonotonic coupling between prestress and stiffness. Increasing cluster number produces collective stiffening when prestressed regions become connected through sufficiently persistent interfaces. At large strain, both solid-like and fluid-like systems approach the corresponding voided-network response as the residual fiber backbone becomes mechanically dominant. Thus, cell-generated prestress controls macroscopic stiffness only together with the organization and persistence of its transmission across the cell-matrix interface.
Read the full abstract
Computer Vision
TempoGround: State-Aware Streaming Visual Grounding with Vision-Language Models
Visual grounding maps language referents to spatial targets and is central to open-vocabulary perception with vision-language models. Existing methods have made substantial progress on single-frame and video-based visual grounding, yet under streaming inputs they still suffer from identity drift, cross-frame inconsistency, and fragile localization under partial occlusion. To address these issues, we present TempoGround, a VLM-native framework that detects cross-frame object correspondence and explicitly models object presence states, thereby enabling accurate and consistent visual grounding under streaming inputs. The key is a curriculum prediction mechanism guided by state-aware cross-frame correspondence: TempoGround resolves 2D instance association, predicts whether each object newly enters, continues in, or leaves the view, decodes the 2D box, and then lifts it to a camera-frame 3D box. As token-level supervision alone cannot capture the geometric objectives of streaming grounding, we further introduce Streaming Grounding Reinforcement (SGR), which optimizes TempoGround with verifiable Grounding, Identity, and Consistency rewards, jointly reinforcing persistent localization and temporally consistent predictions. We carefully design a three-stage training strategy and train TempoGround on large-scale data. We evaluate visual grounding under causally streaming inputs on multiple challenging benchmarks: TempoGround improves F1_2D@0.5 and F1_2D@0.95 by 4.4 and 0.5 on average, and F1_3D@0.25 and AP_3D by 6.2 and 7.5, respectively. These results demonstrate that TempoGround provides a practical foundation for visual grounding under streaming inputs.
Read the full abstract
Robotics
Humanoid Safe Stop via Learned Stoppability Value
Humanoid robots responding to emergency stop commands typically execute a fixed maneuver, without reasoning about whether a safe stop is actually feasible from the current state. We cast emergency stopping as a reach-avoid problem and propose Safe-Stop, a task-agnostic framework that pairs a learned stop policy with learned stoppability estimators. The estimators are complementary: a stop-probability estimator supervised by the actual outcomes of the fixed stop policy, and a reach-avoidance estimator supervised by a Hamilton-Jacobi backup over physical state. The first captures emergent stopping behavior of the learned controller; the second provides a complementary recoverability signal. Because the stop policy and estimators do not depend on the behavior policy that preceded the stop command, they transfer across diverse upstream tasks without retraining. At deployment, the two estimates are combined: Safe-Stop commits to the stop only when both estimators indicate that stopping remains feasible, otherwise it hands off to a fall policy, instantiated as a damping fallback. This agreement check yields decisions that are robust without sacrificing reactivity.
Read the full abstract
Theoretical Economics
Strategic Centrality and the Emergence of Core-Periphery Networks
We study a network formation game in which agents sponsor links at a linear cost in order to maximize centrality, defined as a weighted sum of walk counts with positive and weakly decreasing weights. This class includes Katz Bonacich centrality and total communicability and captures environments in which access decays with distance. Our main result is a sharp equilibrium characterization: every Nash equilibrium network is core periphery, meaning there exists a set $C$ such that every node is linked to every node in $C$, and there are no other edges. The driving force is that linking to better connected agents generates many additional short connections at once, so incentives concentrate links on a dense subset. Finally, we show that the welfare maximizing network is always either empty or complete.
Read the full abstract
High Energy Physics (Phenomenology)
Recoil Geometry Unmasks Gluon Saturation in Forward $Z^0$ Production
Gluon saturation produces characteristic transverse-momentum broadening in nuclei, but QCD radiation largely washes out this signature. We show that fiducial recoil subtraction turns detector acceptance into a transverse-momentum projector that unmasks the broadening in forward $Z^0$ production. Subtracting the hadronic recoil measured in a chosen rapidity interval from the boson transverse momentum defines a residual momentum. At leading power, the radiative recoil in this interval cancels, while the residual momentum retains sensitivity to the small-$x$ nuclear field. Combining a CGC description of the small-$x$ target with soft-collinear effective theory (SCET) resummation for finite rapidity coverage, we find that a benchmark rapidity coverage $|\eta^{\rm lab}|<2.5$ lowers the effective hard scale from $M_Z\simeq 91.2$ GeV to about $7.5~\mathrm{GeV}$ of the Sudakov evolution. Increasing the saturation scale broadens the residual-momentum distribution and weakens recoil alignment, whereas wider coverage makes the proton–nucleus separation clearer in both observables. Detector geometry thus provides tunable control over perturbative recoil, enabling a probe of nonlinear small-$x$ QCD.
Read the full abstract
Statistical Mechanics
Effect of reactive targets on diffusive transport with intermittent restarts
We study search processes in a complex environment using the strategy of stochastic resetting. A common occurrence in these environments is targets with finite reactivity. Stochastic resetting has emerged as a powerful mechanism for optimizing search processes by curtailing long, unproductive excursions inherent to diffusive dynamics. Most theoretical studies, however, assume perfectly absorbing targets – an idealization that overlooks the finite reactivity commonly encountered in realistic chemical and biological systems. In this work, we investigate the interplay between stochastic resetting and finite-target reactivity in reaction - diffusion processes. Considering a one-dimensional system with multiple reactive targets, we demonstrated that the target reactivity modifies the optimization landscape. We uncover distinct regimes in which resetting enhances transport efficiency towards specific targets based on their chemical kinetics. As a consequence, the optimal resetting rate becomes intrinsically reactivity dependent. Our results identify target reactivity as a crucial control parameter governing stochastic transport and provide insights into reaction-diffusion processes in complex media.
Read the full abstract
Game Theory
Fair Stable Matching: A Nash Social Welfare Approach
While traditional stable matching algorithms, such as the Gale-Shapley algorithm, prioritize stability, they may fall short of achieving equitable outcomes among participants. We study the role of Nash social welfare (NSW) as a fairness objective in the classic stable marriage problem. We develop SNSW-Alg that finds a stable matching that maximizes Nash social welfare under rank-induced utilities in $\tilde{\mathcal{O}}(n^4)$ time, where $n$ is the number of men or women. We demonstrate that SNSW-Alg balances equity while preserving stability. We empirically evaluate our methods across diverse preference distributions, demonstrating significant gains in fairness without substantial losses in other key measures such as regret, egalitarian criterion, and sex equality. Our findings suggest that the stable matching produced by SNSW-Alg is statistically Pareto-undominated by stable matchings based on other fairness measures - regret, egalitarian, and sex equality. This study offers compelling insights for designing fair-stable matching.
Read the full abstract
Neural and Evolutionary Computing
LLM-Driven Joint Evolution of Coupled Heuristics Components for Routing Optimization
Heuristic design for combinatorial optimization remains heavily reliant on expert knowledge, while existing large language model (LLM)-enhanced evolutionary methods typically evolve isolated algorithmic components, even when one determines the search state on which another operates. This paper proposes LLM-driven Heuristic Components Joint Generation (LLM-HCJG), a population-based framework that jointly generates and co-evolves interdependent heuristic components under a shared design blueprint. Applied to guided local search (GLS), LLM-HCJG couples solution initialization with penalty construction and embeds the generated pair into an enhanced online search mechanism. The resulting form is further transferred from the traveling salesman problem (TSP) to the capacitated vehicle routing problem (CVRP). Theoretical analysis establishes the non-separable state-transition effects between the two components and the advantage in generation consistency. Across synthetic instances and 41 public TSPLIB/CVRPLIB benchmarks, LLM-HCJG attains consistently low optimality gaps, including best or tied-best results on 28 of 29 TSPLIB instances and all 12 CVRPLIB instances. Ablation and structural analyses further indicate that these gains are associated with cross-component compatibility and alignment rather than isolated-component recombination. These results support effective cross-instance transfer within the evaluated routing settings under limited-sample, modest-cost training.
Read the full abstract
Hardware Architecture
Atlas: Algorithm-Hardware Co-Design for On-Device City-Scale 3D Gaussian Splatting in VR
3D Gaussian splatting (3DGS) has drawn significant attention in the architectural community recently. However, enabling city scale 3DGS on mobile VR devices remains challenging, as the memory requirement of large scale scenes far exceeds the memory capacity of today's mobile GPUs. This paper presents Atlas, an on device city scale 3DGS rendering framework that enables scalable rendering without runtime Internet access. The key insight is that although the full 3DGS model is massive, each frame only requires a small subset of Gaussians under the current pose and level of detail requirement. Based on this insight, Atlas introduces a hierarchical memory offloading mechanism that dynamically loads only necessary Gaussian data into device memory. To further improve performance, Atlas proposes temporal aware LoD search and stereo rasterization to avoid redundant computation in VR. We further show that our technique can be integrated with existing 3DGS accelerators with negligible hardware overhead. Overall, Atlas achieves 18.5x speedup over the GPU baseline and 3.9x speedup over the state of the art 3DGS accelerators, with 92.4% energy savings.
Read the full abstract
Statistical Mechanics
Nonequilibrium corrections to conserved Ising criticality in scalar active matter: Ward identities, spectrum, and long crossovers
We identify the slowest-decaying nonequilibrium perturbations near the three-dimensional conserved Ising critical point and determine their impact on finite-size observables. We study two classes of perturbations: a field-dependent noise-to-mobility ratio $\Theta(\phi)=D(\phi)/M(\phi)$ and the gradient activity of Active Model B+. Starting from the Martin-Siggia-Rose-Janssen-De Dominicis action, we compute the linearized flow using the functional renormalization group. The transport sector is block triangular, with leading odd eigenvalue $y_{\Theta_1}=-\Delta_\phi$, where $\Delta_\phi=(d-2+\eta)/2$. In the gradient sector, removing the detailed-balance direction leaves two genuinely nonequilibrium modes, chemical and current-like. Two smooth regulators give $y_{\Theta_1}\simeq-0.52$, $y_J\simeq-0.56$, and $y_{\rm ch}\simeq-0.89$. A translation Ward identity expresses the current operator as the divergence of the stress tensor. Together with conservation and Itô causality, this forbids chemical operators from generating the current mode, making the nonequilibrium stability matrix triangular; an independent two-loop calculation in $d=4-\varepsilon$ finds no additional current contact counterterm. Within the FRG truncation, $y_J-y_{\Theta_1}=-\eta$; beyond it, this relation requires the absence of an additional contact anomaly. Using the 3D Ising value $\eta=0.0362978(20)$ [Kos et al., 2016] gives $y_{\Theta_1}\simeq-0.5181$ and $y_J\simeq-0.5544$. Because these exponents nearly coincide, single-power fits yield amplitude-dependent apparent exponents and crossover lengths may exceed accessible system sizes. We derive the resulting finite-size scaling rules: odd block observables respond linearly to activity, while even observables receive only quadratic corrections.
Read the full abstract
Computer Vision
LookStep: Efficient Vision-Language Navigation with Linguistic Foresight and Event Driven Memory
Vision-Language Navigation (VLN) requires an embodied agent to follow natural-language instructions in unseen environments. Recent progress has been largely driven by Multimodal Large Language Models (MLLMs). Existing methods follow a next-step action prediction paradigm, supervising only the expert action, which requires a high quantity of data for training. They also rely on cognitive maps, accumulated historical frames, or external 3D tools to maintain states, leading to high computational and memory overhead. To realize resource efficiency VLN, we propose LookStep, a unified end-to-end framework that combines Language Centric Future State Modeling and Event Driven Rolling Memory that uses language labels to generate coarse-grained navigation progress and future states for each candidate action, while autonomously deciding whether to write each observation into a bounded rolling memory with a semantic role. We validate LookStep empirically. On VLN-CE tasks, LookStep outperforms existing methods under the same training settings, achieving a 49.7% success rate on R2R-CE Val-Unseen with better memory efficiency and less data usage. Code and model is available at https://github.com/kunyang-YU/LookStep.
Read the full abstract
Computer Vision
GlyphAnchor: Enhancing Visual Text Rendering via Position-Anchored Glyph Priors
Rendering accurate text remains difficult for image generation and editing models, especially when the target contains long, complex, and densely arranged text or rare characters. Existing approaches either improve native text rendering through stronger backbones and data-centric training without explicit glyph priors, or incorporate glyph priors through specialized designs that remain insufficiently accurate and robust under challenging scenarios. We introduce GlyphAnchor, a novel text-rendering enhancement method for both text-to-image and image-editing diffusion transformer models. GlyphAnchor enhances the backbone with lightweight glyph patch conditions whose positions are anchored to the target image through the model's native positional encoding. We train this capability with staged supervised finetuning and further refine it with text-aware post-training to improve robustness. We also introduce InfoTextBench, a benchmark for evaluating text-rich visual text rendering in both generation and editing settings. Experiments across multiple backbones and benchmarks, including long, complex, and densely arranged text and rare character scenarios, show that GlyphAnchor consistently improves text fidelity while preserving overall image quality.