Peer Revenue Publish or profit

Papers

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

All of arXiv

Friday 4 September

Earth and Planetary Astrophysics

Updated in-flight calibration of the Hayabusa2/NIRS3 spectrometer: new global near-infrared photometric properties of asteroid (162173) Ryugu

The Near InfraRed Spectrometer (NIRS3) onboard Hayabusa2 observed asteroid (162173) Ryugu, searching especially for signatures of water-bearing minerals. However, during the mission, absolute reflectance measured by the Telescopic Optical Navigation Camera (ONC-T) and NIRS3 showed a systematic offset. Additionally, touchdown operations have modified the instrumental response, which has made post-second-touchdown data less reliable. Correction of these issues is crucial to interpret the subtle spectroscopic differences across the surface, better understand the origin and evolution of carbonaceous asteroids, and support the Hayabusa2$\sharp$ mission's future observations of the asteroids (98943) Torifune and 1998 KY26. We used Ryugu observations at different mission phases for relative calibrations and lunar observations for the absolute calibration. Moon spectra acquired during the first Earth swing-by were compared to simulated ones from photometric models based on lunar orbiter and ground-based data. We then applied the updated calibration to the full Hayabusa2 proximity phase dataset, enabling disk-resolved global and regional photometric analyses. We showed that the data obtained following the first and second touchdowns required downscaled corrections of $5.4 \pm 1.8 $% and $10.3 \pm 0.4 $% at 2 $\mu m$, respectively, while the absolute calibration revealed that NIRS3 reflectance was underestimated by about 8% compared with the original calibration. Our new photometric parameters are consistent with previous studies and indicate a weak east-west dichotomy in phase ratio and phase reddening, pointing to higher surface roughness and higher abundance of fine-grained regolith in Ryugu eastern hemisphere.
Read the full abstract

All authors Antonin Wargnier, Eri Tatsumi, Koki Yumoto, Mayumi Ichikawa, Shin-ya Murakami, Yuichiro Nagai, Kazuhiro Honda, Yasuhiro Yokota, Toru Kouyama, Takahiro Iwata, Moe Matsuoka, Satoshi Tanaka, Kohei Kitazato

Mesoscale and Nanoscale Physics

Bias-field control of the Neel skyrmion nonlinearity in a confined nanostructure

We study the nonlinear dynamics of a cylindrical skyrmion-based oscillator within the framework of a generalized Thiele model. We demonstrate the influence of an bias magnetic field applied perpendicular to the plane of the nanocylinder on the oscillation frequency and on the nonlinearity coefficient. It is shown that the field tunes the response frequency, while also producing a substantial change in the nonlinear frequency shift. We find that the nonlinearity coefficient reverses its sign as the field crosses a certain critical value. The variation of the nonlinearity coefficient with the field is clearly illustrated by the observed qualitative changes in nonlinear amplitude-frequency responses. Controlling the nonlinear properties of a skyrmion oscillator by changing the bias magnetic field opens up prospects for creating tunable computational elements for neuromorphic applications.
Read the full abstract

All authors A.V. Valkov, A.A. Matveev, O.Yu. Arkhipova, R.V. Shcherbakov, A.R. Safin, S.A. Nikitov

Machine Learning

BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference

Large Reasoning Models (LRMs) achieve superior problem-solving through extended Chain-of-Thought (CoT) generation, but the resulting key-value (KV) cache grows linearly with sequence length and creates severe memory bottlenecks, often exceeding GPU capacity for long reasoning traces. Existing KV cache compression methods rely on recent queries to estimate future token importance, implicitly assuming these serve as reliable proxies for future attention patterns. We demonstrate that this assumption fails in long-horizon reasoning: certain decoding steps generate Thought Revisiting Tokens (TRT) that re-attend to distant previous context, such as task-solving plans formulated early in the trace. Through systematic analysis, we discover that queries corresponding to the TRT cluster into a small number of similarity groups in the embedding space. Based on this insight, we propose BeaconKV, a training-free KV cache compression method that maintains beacon queries, compact representatives for each global query cluster, to anticipate which KV pairs will be revisited without storing the entire query history. Across four open-source LRMs and diverse reasoning benchmarks, BeaconKV generally outperforms existing compression methods, achieving up to $5.8\times$ memory reduction while nearly preserving full cache accuracy and improving throughput by over $4.3\times$.
Read the full abstract

All authors Janghyeon Kim, Minsoo Kim, Kyuhong Shim, Jungwook Choi

Cryptography and Security

Robust Coverless Linguistic Steganography via Sentence Embedding Space with Global Resynchronization

Linguistic steganography enables covert communication through natural language. Existing methods heavily rely on token-level operations and struggle to maintain reliability under word- and sentence-level textual perturbations. Moreover, variable-length coding-based schemes are highly susceptible to bit-slippage under minor disturbances, as perturbations cause desynchronization between embedded and extracted bit sequences. To address these issues, we propose a robust coverless steganographic framework that operates in the sentence embedding space rather than the token space. Specifically, secret messages are encoded as hierarchical clustering paths in the sentence embedding space, which enhances decoding stability against word- and sentence-level textual perturbations. To tackle the bit-slippage problem, we introduce a Global Resynchronization Mechanism (GRM) that reframes variable-length bitstreams as discrete symbols anchored to semantic subspaces, decoupling local embedding failures from global message recovery. Experimental results demonstrate that under word- and sentence-level perturbations, our approach achieves substantial improvements in robustness, while maintaining effective embedding capacity and exhibiting strong resistance to statistical analysis.
Read the full abstract

All authors Lizhi Xiong, Yuping Lu, Jun Li, Ziqiang Li, Zhangjie Fu

Optimization and Control

Quantitative linear approximation for controllability of quasi-linear parabolic equations

This paper studies the relationship between the controllability of a quasi-linear parabolic equation and that of its linear counterpart. Under suitable assumptions, we prove that for sufficiently small initial data both systems are null controllable, and that the differences between their corresponding controls and states satisfy a higher-order error estimate. Unlike standard controllability problems, the controls constructed here not only steer the system to a prescribed target at a given terminal time, but are also designed to ensure that the deviation between the nonlinear and linear dynamics obeys the aforementioned estimate. This method is applicable to general nonlinear partial differential equations, and the approximation order for controllability is optimal.
Read the full abstract

Signal Processing

Mutual-Coupling-Aware Movable and Fluid Antennas on Holographic Surfaces: A Wavenumber-Domain Circuit-Field Unification

Movable and fluid antenna systems turn antenna position into a design variable. At sub-wavelength spacings, however, their behavior is governed by mutual coupling, modeled today by two disjoint traditions: circuit-theoretic impedance matrices with element-level constants, and field-theoretic kernels with norm-type power constraints. This paper unifies the two. Starting from the impedance kernel of a holographic surface, a Poynting-anchored balance identifies its resistive part with ohmic plus radiated power and its reactive part with stored-energy imbalance, and a circuit-field equivalence shows that the multiport impedance matrix is the kernel sampled at the port separations, in a single closed spherical-Hankel form. In the wavenumber domain the resistive kernel asymptotically diagonalizes in the aperture size: visible modes radiate at closed-form prices, evanescent modes only dissipate, and a flexible port becomes a constant-modulus codeword whose coupling is the pullback of the spectral weight. Coupling-aware multi-user sum-rate maximization over precoders and port positions is then formulated under physical power and voltage constraints and solved by weighted-MMSE and projected-gradient steps with closed-form gradients. A modal relaxation upper-bounds every port configuration and seeds the search by FFT-based codeword projection. A half-wavelength corollary and a superdirectivity margin quantify when coupling hurts, and when it helps.
Read the full abstract

Astrophysics of Galaxies

Rubin J122659.4+090236: An Extremely Low Surface Brightness Galaxy Candidate Discovered in the Rubin LSST Early Data Preview 2

We report the serendipitous discovery of an exceptionally low surface brightness galaxy (LSBG) candidate, Rubin J122659.4+090236, in Rubin Observatory imaging of the interacting NGC 4410 system, identified in the Cosmic Treasure Chest public release. 2D Sérsic modelling of the Rubin g, r, and i images reveals a nearly round system with a shallow profile (n ~ 0.4), an effective radius of R_e ~ 6”, and central surface brightnesses of $\mu_{0,g}=27.52\pm0.04$, $\mu_{0,r}=27.62\pm0.07$, and $\mu_{0,i}=27.04\pm0.08$ mag arcsec$^{-2}$. EAZY photo-z fitting favours an intermediate-z solution at z~0.3, while a low-redshift solution at z~0.028, consistent with the NGC 4410 system, is also permitted by a restricted EAZY fit over 0<z<0.1 without imposing a redshift prior. These alternatives imply substantially different physical interpretations, ranging from a diffuse dwarf-like system to an exceptionally extended background LSBG. This discovery demonstrates Rubin's sensitivity to extremely diffuse galaxies and highlights the potential of the LSST survey to uncover large samples of such elusive systems across wide areas, enabling systematic studies of the LSBG population and its role in galaxy evolution.
Read the full abstract

All authors Dipanjan Mitra, Kanak Saha, Sugata Kaviraj, Alister W. Graham, Sandro Tacchella, Ricardo Demarco, Dieu D. Nguyen, Elham Saremi, hernandez toledo, hector., Jose Benavides, Natanael M. Cardoso, Samuel Boissier, Dan S. Taranu, Bililign Dullo, Patricia B. Tissera, Matias Blana, Hao Fu, Priyanka Chakraborty, Kevin A. Pimbblet, Tayyaba Zafar, Anirban Dutta, Aidan P Cotter, Darko Donevski, Katarzyna Malek, Rossella Ragusa, Agnieszka Pollo, Unnikrishnan Sureshkumar, Aritra Ghosh, Antonio Vanzanella, Hareesh Thuruthipilly

General Relativity and Quantum Cosmology

Static Spherically Symmetric Solutions in Modified Entropic Gravity

Within the entropic gravity paradigm introduced by Verlinde, one assumes that the microscopic degrees of freedom residing on the holographic screen obey the equipartition law of energy. Nevertheless, implications of statistical mechanics suggest that this energy sharing can acquire corrections that depend on temperature. Taking such modifications into account leads to altered gravitational field equations when derived from thermodynamic considerations. We solve the resulting modified Einstein equations in the case of a static, spherically symmetric spacetime and determine the general structure of the metric components. Our findings indicate that if the temperature correction function behaves as $f(T)\propto T^{2}$, the corresponding spacetime geometry reduces to the flat Minkowski metric. Therefore, to obtain small deviations from flatness, it is necessary to consider slight departures from this purely quadratic temperature dependence. Such deviations can be interpreted as encoding additional gravitational effects, potentially associated with matter contributions that are not entirely described by the holographic screen. As a result of these effects, the effective gravitational potential acquires a logarithmic correction, which can give rise to deviations from Newtonian gravity. This aspect is especially significant in the regime of very weak gravitational fields, where these corrections may affect particle motion and could carry implications for both astrophysical and cosmological contexts.
Read the full abstract

All authors A. Rostami, M. Rostampour, A. Yarahmadi, K. Rezazadeh

Computer Vision

ARC-Loc: Leveraging Azimuthal Ray Convergence as a Geometric Cue for Direct Cross-View Localization

Cross-view localization (CVL) estimates the pose of a ground image by matching it to a geo-referenced satellite image. To bridge the extreme viewpoint gap, mainstream pipelines rely on Bird's-Eye-View (BEV) transformations or 2D-to-3D lifting. However, deriving 3D structures from a single ground image is fundamentally ill-posed, causing these methods to endure geometric distortions and computational costs during 3D lifting or BEV projection. Furthermore, relying on external depth foundation models to resolve this introduces latency and remains susceptible to noisy predictions. In this work, we present a different approach inspired by a human navigation technique called resection, that can perform direct ground to satellite image matching and localization without relying on external depth foundation models. The key insights of our method are that (i) ground keypoints can be translated into azimuthal rays on the satellite map, and (ii) these rays ideally converge at the user location. Exploiting this geometric constraint through direct line-to-point correspondences, we introduce a minimal Azimuthal Ray Convergence (ARC) solver to identify the intersection, alongside an ARC loss to optimize the matching network. By eliminating dependencies on computationally heavy BEV transformations and external depth foundation models, our approach achieves faster, memory-efficient inference, while its explicit feature matching ensures straightforward compatibility with existing frameworks. Experiments on VIGOR and KITTI demonstrate that ARC-Loc maintains competitive localization accuracy compared to recent approaches, highlighting its practicality.
Read the full abstract

All authors Hyeongsik Kim, Mincheol Kim, Heejoon Moon, Je Hyeong Hong

High Energy Physics (Phenomenology)

Tracing Gluon Saturation through Hadronization at EIC

Gluon saturation provides a window into the nonlinear nature of the strong interaction in nuclear matter. One direct consequence of saturation is the $\boldsymbol{k}_T$ broadening in the final state.We investigate how hadronization reshapes conventional signatures of gluon saturation at EIC within a complete event-generator framework. To this end, we implement in eHIJING an initial-state-radiation algorithm based on nonlinear small-$x$ evolution and complete the events with beam remnants, final-state radiation, and hadronization. In our simulations, the signal of parton-level nuclear $\boldsymbol{k}_T$ broadening is strongly diluted in both the nucleon energy correlator and leading-dihadron azimuthal decorrelation after hadronization. Global hadronic recoil, by contrast, remains sensitive to the underlying $\boldsymbol{k}_T$ broadening. We further demonstrate that Bayesian unfolding of the global hadronic recoil provides access to the underlying hard-scattering $\boldsymbol{k}_T$ distribution. These results establish the global hadronic recoil as a promising saturation observable at the EIC.
Read the full abstract

All authors Weiyao Ke, Xu Liu, Yu Shi, Xin-Nian Wang, Jian Zhou

Machine Learning

Fractal basins trap latent reasoning

Reasoning allows artificial intelligence models to revisit and correct their mistakes, enabling recent frontier advances in mathematical theorem solving, software engineering, and autonomous task planning. Reasoning models are widely observed to reason for longer on harder tasks, but the general mechanism responsible for these slowdowns is unknown. Here, we show that reasoning models exhibit transient chaos, a physical consequence of the computational complexity of difficult tasks. As a consequence, we show that diverse leading reasoning models are dynamical systems with fractal basins, with fractality increasing with task difficulty across diverse tasks like Sudoku and maze solving, visual puzzles, and mathematical logic. We show that transient chaos emerges due to reasoning becoming trapped for extended durations near saddle points, which we show correspond to nearly-correct attempted solutions of the underlying problem. Our results show that reasoning slowdowns are an inevitable consequence of problem hardness in modern artificial intelligence models, and establish reasoning traces as a rich new class of dynamical system.
Read the full abstract

All authors Jeffrey Lai, Anthony Bao, John Quinn, William Gilpin

Artificial Intelligence

Why We Care About Understanding: Competence through Predictive Compression

What is the relation between understanding and compression, and why does human understanding take such a heavily compressed form? Across information theory, machine learning, and AI research, a substantial tradition identifies understanding with compression-a thought captured in Gregory Chaitin's dictum that "comprehension is compression." Philosophers, by contrast, have characterized understanding in terms of grasping connections, giving explanations, and handling novelty. This paper bridges the two pictures through three interlocking theses. The first concerns the concept of understanding: it serves as an efficient proxy for a distinctive form of robust competence, enabling us to identify whom to trust and whom to learn from. The second concerns the state of understanding: to understand a domain is to possess a mental model of its relational structure that enables prediction, and what enables prediction enables compression, because what becomes predictable need not be stored separately. Compression is therefore not identical with comprehension, but its representational shadow. The third concerns the characteristically human form of understanding: the fiduciary and transmission functions highlighted by the first thesis impose pressures of demonstrability and transmissibility that drive human understanding toward principled simplicity. The resulting framework explains both the appeal and the limits of compressionist accounts of understanding while shedding light on the inscrutability of AI systems.
Read the full abstract

Information Retrieval

SAM-D2Q: Aligning Multimodal Doc2Query with Search Demand and Conversion for E-commerce

E-commerce search often suffers from vocabulary mismatch between user queries and merchant-authored product titles, since short titles cannot fully cover diverse user expressions or visual product attributes. Although Doc2Query alleviates this issue by generating pseudo-queries for document expansion, traditional methods are text-only and not optimized for e-commerce business objectives. As a result, they may produce semantically plausible but commercially ineffective expansions and miss key attributes present in product images. To this end, we propose E-commerce Search-Aligned Multimodal Doc2Query (SAM-D2Q), a business-aligned multimodal document expansion framework for e-commerce search under Boolean retrieval constraints. SAM-D2Q consists of three stages: (1) task-adapted multimodal supervised fine-tuning to enhance vision-language understanding of product titles, images, and user queries; (2) multimodal data augmentation to improve perception of key visual attributes and expansion coverage; and (3) reinforcement-learning-based preference alignment toward search business objectives, encouraging the model to generate pseudo-queries that better match user intent and commercial value. Offline experiments show that SAM-D2Q substantially improves retrieval performance over traditional Doc2Query methods. Deployed in the AliExpress production search system, SAM-D2Q improves online business metrics, increasing GMV by +3.38% and Pay Count by +2.27%.
Read the full abstract

All authors Hui Zhou, Jian Hui Ji, Lei Ma, Rong Xiao, Xiaoyi Zeng

Statistical Methodology

Finite Mixtures of Generalized Estimating Equations for Clustering Multivariate Correlated Outcomes

Multivariate correlated outcomes occur across disciplines, including ecology, social sciences, and psychometrics. This paper focuses on clustering these outcomes across observational units, specifically, finding groups of units with the same “outcome profile". Our motivation comes from bioregionalization in ecology, which aims to cluster sites into bioregions with the same species profiles, where site membership can depend on environmental or habitat covariates. To accomplish this, we propose finite mixtures of generalized estimating equations (MixGEE). Unlike existing approaches to model-based bioregionalization, MixGEE partitions sites into regions while accounting for between-species correlations through a region-specific working correlation structure. Thus, each region is characterized by a marginal species mean vector and a between-species correlation matrix. Unlike likelihood-based finite mixture models, MixGEE does not require a full joint distribution of the multivariate outcomes. Instead, we construct a pseudo-posterior probability for region membership motivated by the large-sample distribution of the estimating equation. This leads to an iterative algorithm alternating between updating these probabilities and solving weighted estimating equations. We determine the number of regions using cross-validation based on predictive performance for held-out sites and species components, and use a clustered Dirichlet random-weight bootstrap for uncertainty quantification. Simulations demonstrate reliable estimation and inference under various correlation structures and more stable selection of the number of groups than methods that ignore dependence. Applying MixGEE to presence–absence records of fish species around the Kerguelen Plateau reveals three distinct fish assemblage profiles with heterogeneous occurrence and within-site correlation patterns.
Read the full abstract

Computation and Language

Discourse Dependency: A Continuous Criterion for Translation Difficulty

Recent calls for harder machine translation benchmarks have not clarified what difficulty should mean. We argue that one meaningful and currently unmeasured axis is referential reach, the distance a segment must look back into its document to resolve the entities and pronouns it contains. We formalize this as discourse dependency (DDP), a metric-free, source-side measure computed from named entity re-mentions and pronominal coreference. Validated against gold coreference, DDP errs one-sidedly in 99.2% of segments, so a high-DDP segment is certified to require long-range context. Applying DDP to WMT24++ and WMT25 shows that both are heavily skewed toward low-DDP segments, which domain labels do not distinguish. Building on DDP, we compare five context injection strategies in an English-Korean post-editing setup, varying context size and selection. As DDP grows, no strategy keeps pace with human post-editing. On segments with DDP >= 15 raters prefer human translations, while automatic metrics register no difference. As frontier systems saturate aggregate scores, DDP shifts evaluation from how well models score to how far they can reach.
Read the full abstract

Computer Vision

MINT: A Unified Model for World-Space Camera and Hand Motion Estimation from Scalable Egocentric Pipeline Supervision

Recovering camera and hand motion in world coordinates from egocentric video is a key capability for activity understanding, robot learning, and augmented reality. Existing systems typically decompose this problem into separate stages for camera motion, depth estimation, hand reconstruction, and trajectory refinement, resulting in substantial computational overhead and preventing the joint modeling of camera and hand motion. We introduce MINT (Minting IN-the-Wild Trajectories), a foundation model for world-space hand motion reconstruction from ego-centric RGB video. From a single shared spatiotemporal video representation, MINT jointly predicts the camera trajectory, field of view (FoV), camera-frame hand states, and per-frame hand observability, and then produces world-space hand motion via explicit coordinate transformations. Training such a model at scale is challenging, since paired world-space camera and hand annotations are scarce. We therefore develop an open-source labeling EGOPIPELINE that converts large collections of public egocentric videos into structured camera-and-hand trajectory supervision. MINT is first pretrained on these large-scale pseudo-labels and then fine-tuned on a small set of high-quality camera-and-hand annotations. Across public benchmarks MINT approaches state-of-the-art accuracy without seeing either benchmark in training, reaching 0.945 frame accuracy, 13.646 mm PA-MPJPE-p and 55.058 px EPE-p for camera-frame bimanual reconstruction on HOT3D, 4.690 mm RPE-T and 0.284 degrees RPE-R for camera trajectory, and a 3.67x end-to-end speedup over the labeling pipeline that supervises it. We release the model, training and inference code, labeling pipeline, and a curated 1,021-hour egocentric trajectory dataset.
Read the full abstract

All authors Zijie Zhu, Weiren Cai, Yizhou Wang, Zhenjie Yang, Yide Liu, Jiahao Chen, Guanqi He

Statistical Mechanics

Universal Thermodynamic Law Governing Stochastic Pendulum Clocks

Clocks' performance, especially their precision, is fundamentally limited by thermodynamic laws as they are physical devices. Yet, it is known that the established thermodynamic uncertainty relation (TUR), which imposes an upper bound on the uncertainty product of the precision of oscillations and entropy production, is violated for underdamped systems such as stochastic pendulum clocks. Here, we show that for a general class of stochastic pendulum clocks described as a weakly nonlinear oscillator the uncertainty product of the phase of a pendulum clock and entropy production takes a universal and simple form that depends solely on the degree of nonlinearity. Its validity and limitations are examined using several representative models of pendulum clocks. Our framework reveals a universal thermodynamic principle governing stochastic pendulum clocks beyond the conventional TUR and provides a foundation for designing optimal pendulum clocks that operate efficiently in stochastic environments.
Read the full abstract

Combinatorics

The Erdős-Pósa Property for Colorful Minors

A colorful graph relation enhances the minor relation by merging color sets along contractions and by allowing the removal of colors; it generalizes rooted minors and models problems on graphs with several, possibly overlapping, annotated vertex sets. A graph has the Erdős-Pósa property for minors if and only if it is planar, by a classical theorem of Robertson and Seymour. In this work we determine, for the colorful minor relation, exactly which colorful graphs have the Erdős-Pósa property. Our characterization takes three equivalent forms. The first is structural: the colorful graphs with the property are those that can be drawn with all their colored vertices on one face and whose colors are, in a precise sense, laid out along that face without interleaving. The second is given by an obstruction set: they are those excluding every member of an explicit infinite family $\mathcal{O},$ of which only $\mathbf{O}(|I|^{4})$ members have colors that are a subset of $I,$ for every finite set $I$ of colors. The third is grid-like: they are exactly the colorful minors of unions of particular families of segregated grids, the colorful analogues of the grids that drive the classical proof.
Read the full abstract

Materials Science

Nuclear quantum effects in the thermal conductivity of solid hydrogen

We use a combination of path integral and lattice dynamics methods to calculate the thermal conductivity of solid parahydrogen. Path integral molecular dynamics is first used to calculate a centroid potential of mean force at each temperature, on which a harmonic phonon basis is constructed using the finite displacement method. We then calculate anharmonic force constants on the centroid potential and solve the linearised phonon Boltzmann transport equation for the thermal conductivity. The resulting renormalised phonon dispersion curves, heat capacity, and thermal conductivity are all in remarkably good agreement with experimental measurements. We find that the heat transport is dominated by collective relaxon modes at low temperatures and cannot be captured by making the phonon relaxation time approximation. Indeed, quantitative agreement with the experimental thermal conductivity is only achieved when the nuclear quantum effects in the centroid potential of mean force, phonon boundary scattering, and hydrodynamic transport are all explicitly considered. This is also likely to be the case for other quantum crystals.
Read the full abstract

Mesoscale and Nanoscale Physics

Nonreciprocal plasmonic response of drift-biased two-dimensional metals

We develop a nonlocal electrodynamic framework for drift-biased two-dimensional electron gases (2DEGs) with Dirac (linear) and parabolic (quadratic) dispersions, deriving closed-form drift-dependent conductivity tensors from the Boltzmann transport equation. We obtain the plasmon dispersion and near-field emission of a point dipole, revealing nonreciprocal propagation in both systems. At equal drift parameter, the parabolic 2DEG exhibits stronger nonreciprocity than the Dirac system, although we find that a larger drift parameter does not by itself produce stronger nonreciprocity: the response is set by the interplay of drift and plasmon nonlocality. As such, the combination of strong nonlocality and experimentally accessible drift velocities of semiconductor 2DEGs identifies parabolic systems as a promising platform for tunable, nonreciprocal plasmonics.
Read the full abstract

Dynamical Systems

Cone upgrade for effective equidistribution and weighted Khintchine–Schmidt theorem on fractals

We prove a fractal analogue of the weighted theorems of Khintchine and Schmidt for products of self-similar measures on the real line whose defining iterated function systems share a common contraction ratio. The proof translates the counting problem into homogeneous dynamics and builds on recent work of Bénard, He and Zhang, who treated the unweighted case. Our main new ingredient is a cone upgrade: an effective double equidistribution theorem for translates of product self-similar measures, uniform over diagonal elements whose weights range over a fixed cone in the positive Weyl chamber, deduced from effective equidistribution for a single weight. This yields a partial effective form of a theorem of Khalil, Luethi and Weiss.
Read the full abstract

Distributed Computing

GreenPipe: Power Modeling for Containerized DNN Inference on Kubernetes Edge Nodes

Distributed DNN inference is increasingly deployed in containerized edge-cloud environments, where workloads run on-device or are exposed to remote clients over the network. Accurate online power estimation on resource-constrained ARM nodes without hardware power counters such as RAPL remains a challenge, and CPU-only models fail to capture multi-resource behavior. We present GreenPipe, an automated profiling-training-validation pipeline that builds multi-resource regression models from external power meter measurements and attributes power to containers proportionally. GreenPipe is evaluated on a Raspberry Pi 4 edge node in a K3s edge-cloud testbed, covering DNN inference with three vision models, multiple precisions, thread counts, and both local and serving scenarios. System-level MAPE is 6.3-9.4%, improving over CPU-stress and utilization-only baselines by 26.9% MAPE on average. We jointly report inference latency and energy per inference, exposing performance-energy trade-offs across workload configurations.
Read the full abstract

All authors Mengxue Wang, Peini Liu, Amir Taherkordi, Jordi Guitart

Mesoscale and Nanoscale Physics

Cascade spin dynamics of excitons localized in indirect-band-gap (In,Al)As/AlAs quantum dots with type-I band alignment

We investigate the spin dynamics of excitons localized in type I (In,Al)As/AlAs quantum dots with an indirect in momentum space band structure. Polarized selective photoluminescence spectroscopy, i.e. fluorescence line narrowing, under magnetic fields up to 5 T applied in the Faraday geometry is employed. The experiment reveals a cascade spin evolution process of excitons in the indirect band-gap quantum dots: an initial short term spin dynamics associated with excited direct exciton states possessing a large oscillator strength is followed by electron relaxation into the X valley of the Brillouin zone and subsequent long term spin dynamics of indirect excitons. The two step mechanism manifests itself in the distinct features of the magnetic field dependences of photoluminescence: two component recovery of optical orientation, two component linear to circular polarization conversion and the presence of the linear polarization plane rotation. At the same time, suppression of the optical alignment shows one-component behavior governed by the spin dynamics of the indirect exciton states. Within the pseudospin formalism, we derive analytical expressions that quantitatively describe the observed dependences and yield estimates for the anisotropic exchange splitting: 210 {\mu}eV for direct excitons and 1.3 {\mu}eV for indirect excitons. Further analysis using the density matrix formalism agrees well with the pseudospin model calculations and shows that the finite optical orientation at zero magnetic field is due to comparable magnitudes of the anisotropic splitting of the indirect exciton states and the splitting of the X-valley electron states caused by the hyperfine interaction with nuclei.
Read the full abstract

All authors S. V. Nekrasov, I. V. Kalitukha, A. A. Golovatenko, Ya. A. Kuznetsova, N. O. Mikhailenko, M. D. Ragoza, T. S. Shamirzaev, Yu. G. Kusrayev

Discrete Mathematics

The fourth generalized Davenport constant of $C_5^3$

For a finite abelian group $G$ and $k \geq 1$, the generalized Davenport constant $D_k(G)$ is the least $\ell$ such that every sequence over $G$ of length at least $\ell$ has $k$ pairwise disjoint nonempty zero-sum subsequences. A theorem of Freeze and Schmid gives $D_k(C_5^3) \geq 5k+10$ for every $k \geq 2$. We prove the matching upper bound: $D_4(C_5^3)=30$, and hence $D_k(C_5^3)=5k+10$ for every $k \geq 2$, so the Freeze–Schmid bound is attained by $C_5^3$ from $k=2$ onward, as it is by $C_2^3$ and unlike $C_3^3$. The proof is finite and computer-assisted. The remaining case reduces to showing that every zero-sum sequence of length $31$ over $C_5^3$ contains a nonempty zero-sum subsequence of length at most five. A saturation argument confines the multiplicities of a hypothetical counterexample to $\{1,2,4\}$, its support pattern to one of $60$ solutions of two linear equations, and its geometry to one of $78$ rank/plane branches normalized to a standard basis; an exhaustive search exhausts every branch with no survivor. The search was carried out by three independently written implementations, and the branch cover was regenerated by separate programs from the lemmas alone; two further machine-verified values, $D_3(C_5^3)=25$ and $s_{\leq 6}(C_5^3)=24$, enter the second statement, and their records accompany the paper.
Read the full abstract

Applied Physics

Noise-Resilient Detection of Neuronal Spikes by a Hopf-Bifurcation Device

Detecting weak transient signals in noise is a persistent challenge in sensing, communication, and electrophysiology. Here, we demonstrate a physical weak-signal detector based on a semiconductor negative differential resistance (NDR) device operated near a Hopf bifurcation. A coherent input that persists over the response time of the dynamical system can drive a transition between quiescent and oscillatory states, whereas faster stochastic fluctuations are largely suppressed. This nonlinear transformation converts a weak analog threshold crossing into all-or-none voltage spikes and therefore provides asynchronous signal detection without a reference clock. Using a modulated photovoltaic signal, we detect a weak frequency component of 100 Hz as a demonstration, at an input signal-to-noise amplitude ratio as low as 1/500, and reliably recover it in the output spectrum. We then apply the same principle to neuronal multisite extracellular recordings. After standard band-pass filtering, the raw microelectrode signal is transformed by the NDR dynamics, enhancing the distinction between neuronal spikes and background fluctuations. The resulting detected spike times agree closely with those obtained using a traditional spike-detection pipeline. These results establish bifurcation-engineered NDR dynamics as a compact hardware approach to noise-resilient signal discrimination and event-based analog-to-digital conversion that will be useful for neuroprosthetic devices.
Read the full abstract

All authors Jitendra Kumar, Roberto Fenollosa, Gonzalo Rivera-Sierra, So-Yeon Kim, Adam Armada-Moreira, Juan Bisquert, Michele Giugliano

Computer Vision

VICAL: Vicinal Consistency Alignment for Long-Tailed Visual Recognition

Multi-expert models have become the dominant paradigm for long-tailed learning, largely attributed to their presumed ability to benefit from expert diversity. However, we revisit this central assumption and reveal that diversity induced by logit adjustment or explicit regularizers does not guarantee better ensemble accuracy. Our work suggests that multi-expert models benefit more from variance reduction than diversity maximization. We introduce VICAL, a VIcinal Consistency ALignment framework that improves long-tailed recognition not by enforcing expert diversity, but by reducing prediction variance. Specifically, our approach comprises two key components: Self-Consistency Learning and Deep Ensemble Distillation. Self-Consistency Learning discourages reliance on unstable high-frequency information, smoothing the local loss landscape and mitigating overfitting, especially for tail classes. Deep Ensemble Distillation promotes cross-expert low-frequency semantic agreement using a low-resolution view, thereby sidestepping optimization conflicts with established knowledge. Extensive experiments on CIFAR-LT, ImageNet-LT, and iNaturalist 2018 show that VICAL consistently outperforms state-of-the-art methods, validating the effectiveness of our consistency-driven design. Our code is available at \href{https://github.com/FlamieZhu/Vicinal-Consistency-Alignment}{VICAL}.
Read the full abstract

All authors Jiangang Zhu, Zheng Wang, Bin Zhu, Yi-Ping Phoebe Chen, Jingjing Chen

Computer Vision

MCPO: Modality-Contrastive Preference Optimization for Multimodal Chain-of-Thought Compression

Recently, multimodal large-scale reasoning models have demonstrated remarkable capabilities in solving complex tasks through long Chains-of-Thought (M-CoT). However, excessively long reasoning trajectories incur substantial computational costs and significant KV-cache pressure. Existing CoT compression and alignment paradigms mainly rely on static rules or single-dimensional preferences, lacking fine-grained cross-modal constraints; as a result, they are prone to inducing visual laziness and hallucinatory reasoning. To address these issues, we propose Modality-Contrastive Preference Optimization (MCPO), a highly sample-efficient two-stage length-compression method that requires fewer than 900 training samples. In the compression stage, we introduce a step-level Normalized Cross-Modal Mutual Information (NCMI) pruning algorithm, which automatically identifies and removes visual-independent reasoning steps by comparing the reasoning discrepancies between with-image and no-image contexts. This significantly reduces redundancy and hallucinatory content in the reasoning chains. In the alignment stage, the model first undergoes supervised fine-tuning to achieve domain-adaptive initialization, followed by optimization using an asymmetric multimodal length-controlled preference loss. This objective adopts a highly nonlinear odds-ratio formulation that provides steep gradients in the with-image context to reinforce length constraints for preferred trajectories, while applying a scaled, flat-gradient linear difference in the no-image context to maintain modality consistency, thereby achieving stable cross-modal preference alignment. Extensive experiments on mainstream base models such as Qwen3-VL-Thinking show that our method can reduce CoT length by up to 69.5% and achieve up to 3.34x end-to-end inference speedup while preserving original accuracy.
Read the full abstract

All authors Guangheng Yang, Zhenliang Ni, Zhenkai Wu, Han Shu, Juan Feng, Wenming Yang, Jie Hu

Earth and Planetary Astrophysics

Mapping the WISPIT2 Planet-Hosting Cavity at Sub-Hill-Radius scales

We present new deep, very high angular resolution observations of the sub-mm continuum (19x13 mas) and 12CO (67x52 mas) emission from the disk surrounding the planet host WISPIT2. Our analysis reveals a clear gas gap carved by WISPIT 2b and an empty cavity within the orbit of WISPIT 2c. We also find a gaseous and optically thin dusty ring located between the orbits of the two planets. Channel maps show a pronounced change in the emission surface height of the disk, which transitions from flat within the cavity to strongly flared beyond the orbit of planet b, producing clear breaks in the velocity channels. A pronounced thunderbolt-shaped feature, identified as a kinematic planetary signature (KPS), is detected at the location of WISPIT 2b, in agreement with predictions from planet-disk interaction simulations. This constitutes the first system with a confirmed kinematic signature associated with a directly imaged planet. By mirroring opposite velocity-channel maps about the systemic velocity, we further investigate the KPS and identify spatially resolved excess emission in the vicinity of WISPIT 2b. We find that this emission is resolved across the full Hill sphere of the planet, significantly exceeding the expected extent of a circumplanetary disk. We therefore suggest that the observed emission arises from a combination of processes, including prominent planet-driven spiral wakes, localized heating induced by the planet and a circumplanetary disk. Although higher spectral resolution is required to disentangle these contributions, our observations demonstrate the potential to directly characterize the birth environment of forming massive planets at scales down to sub-Hill radius.
Read the full abstract

All authors Myriam Benisty, Stefano Facchini, Richard Teague, Francesco Zagaria, Pietro Curone, Viviana Pezzotta, Jaehan Bae, Gabriele Cugno, Daniel J. Price

Signal Processing

Task-oriented Framework for Communication-Efficient Federated Learning: From Isolated Optimization to Holistic Synergy

Communication bottlenecks remain a primary obstacle to the large-scale deployment of federated learning (FL). This article proposes a comprehensive framework for building communication-efficient FL, founded on three fundamental pillars: model compression, client selection, and resource allocation. We first survey state-of-the-art techniques for each pillar, specifically elucidating how quantization, pruning, and low-rank approximation reduce payloads; how intelligent client schedulers exploit heterogeneity; and how emerging communication paradigms such as Integrated Sensing and Communication (ISAC) and Over-the-Air Computation (AirComp) redefine bandwidth and energy utilization. Subsequently, these insights are unified through a task-oriented design philosophy that couples strategy selection with cross-layer, multi-objective optimization. To validate the proposed framework, we present an autonomous driving case study with two complementary experiments: a task-oriented client scheduling strategy that improves object detection accuracy under the same communication time budget, and a joint quantization-bandwidth optimization that further reduces total training time under dynamic networks. Together, the experiments demonstrate the advantages of holistic task-oriented design for real-world FL deployment.
Read the full abstract

All authors Fuqiang Pan, Yan Liu, Erwu Liu, Arumugam Nallanathan

Earth and Planetary Astrophysics

The metallicity of giant exoplanets around low-mass stars and how Ariel can help

Giant planets orbiting low-mass stars represent a unique population of giant planets that can be studied to constrain planet formation theory. Surveys of transiting giant exoplanets around M-dwarfs (GEMS) allow for measurement of their masses and radii, which in turn can be used to estimate their bulk densities. Coupling these observations to interior structure and evolution models can yield the bulk metallicity of these planets, however there are degeneracies to this that are improved by measurements of atmospheric metallicity. Estimates of bulk metallicity can be crucial to our understanding of planet formation timescales and mass budgets, particularly in these extreme (planet-to-star) mass ratio systems. Here we show that GEMS are expected to have a low bulk metallicity as a natural result of the planet formation process, and the low efficiency of planetesimal capture post-formation. To test this empirically, we need a larger sample of GEMS with atmospheric measurements, which can reduce some of the degeneracies with the interior modelling of planets. Measuring the atmospheric composition of GEMS with JWST and Ariel will be crucial for better understanding this unique planetary type, which sits at the tail of standard planet formation conditions.
Read the full abstract

All authors Ravit Helled, Sho Shibata, Shubham Kanodia, Billy Edwards