Peer Revenue Past citations are no guarantee of future returns

Papers

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

All of arXiv

Monday 14 September

Functional Analysis

The Strong Naor–Schechtman Convolution Inequality on the Product of Cyclic Groups

In this paper, we resolve the convolution inequality problem posed by Naor and Schechtman in \cite[Question 6.1]{N-S2016} (see also \cite{Na2016}). More precisely, we establish the following strong convolution inequality on products of finite cyclic groups: For each $1\leq p<\infty$ there exists $c_{p}>0$ such that for every $n$, $m\in \mathbb{N}$ the following holds \[ \begin{split} &\sum_{\varepsilon\in \{-1,1\}^{n}}\sum_{x\in \Z}\left|E_{\{1,\cdots,n\}}f(x+\varepsilon)-E_{\{1,\cdots,n\}}f(x-\varepsilon)\right|^{p}\\ \leq& c^{p}_{p}\sum_{\varepsilon\in \{-1,1\}^{n}}\sum_{x\in \Z}\left|\varepsilon_{j}\left[E_{\{1,\cdots,n\}\setminus\{j\}}f(x+e_{j})-E_{\{1,\cdots,n\}\setminus\{j\}}f(x-e_{j})\right]\right|^{p}, \end{split} \] $E_{A}f(x)=\frac{1}{2^{n}}\sum_{\delta\in \{-1,1\}^{n}}f\left(x+\sum_{j\in A}\delta_{j}e_{j}\right)$ for every $A\subseteq \{1,\cdots,n\}$. The growth order we obtain here $c_{p}=O(p)$ as $p\to \infty$. For the endpoint $p=\infty$, we show by a counterexample that the strong convolution inequality fails for the $L_{\infty}$, and we prove that the appropriate endpoint for $p=\infty$ is the martingale $BMO$ space.
Read the full abstract

All authors Yong Jiao, Sijie Luo, Dmitriy Zanin, Dejian Zhou

Numerical Analysis

Nested QMC designs on spheres

Nested cubature rules, in which each refinement retains all previously used nodes and thus reuses earlier function evaluations, are natural in multilevel and adaptive integration. For every fixed $s>d/2$, we show that the equal-weight QMC integration rate on $\mathbb S^d$ is compatible with such nested point sets. In the subcritical range $d/2<s<d$, cumulative unions of geometrically growing QMC blocks yield nested QMC design sequences whose successive cardinality ratios converge to any prescribed $\rho>1$. At and above the critical index $s=d$, where the block-averaging estimate no longer yields the optimal rate, we prove an equal-weight completion theorem based on low-frequency discrepancy cancellation. A block-sensitive estimate sharpens the iteration and yields nested QMC designs for all $s\ge d$ with $N_{j+1}\lesssim(j+1)^{2s/d-1}N_j$. Every prescribed finite point set also admits an optimal-rate completion.
Read the full abstract

Materials Science

Freestanding Antiferromagnetic Oxide Membranes: Synthesis and Characterization of Cr$_2$O$_3$

Antiferromagnetic (AFM) insulators stand out as a promising class of materials for fast switching, energy-efficient magnon-based technology, but our understanding of how to tune these properties with the lattice remains limited. As an AFM insulator with strong magnetoelectric response, Cr$_2$O$_3$ membranes stand out as a promising platform to explore multi-modal tunability via strain. However, fabricating these membranes is challenging due to the scarcity of etchable sacrificial layers with compatible lattice constants and sufficient surface quality. We address this issue by utilizing La$_{0.7}$Sr$_{0.3}$MnO$_3$ (LSMO) as a sacrificial layer, enabling successful fabrication of millimeter-scale Cr$_2$O$_3$ membranes. Substrate-free characterization reveals that strain due to the lattice mismatch during growth is released by the formation of small polycrystalline domains, preserving single-crystalline order over ~90% of the membrane area. Bulk-like structural properties are confirmed by transmission electron microscopy, X-ray diffraction, Raman spectroscopy, and second-harmonic generation spectroscopy. Platinum Hall measurements reveal an above-room-temperature N$\'e$el transition. These results establish Cr$_2$O$_3$ membranes as a viable platform for strain-tuning antiferromagnetism and magnetoelectricity.
Read the full abstract

All authors Ella Blake, Tiffany C. Wang, Yi Cui, Minyong Han, Yinchuan Lv, Christopher T. Parzyck, Octave Duros, Yan Che, Matthias C. Hoffmann, Wei-Sheng Lee, Yi Cui, Harold Y. Hwang

Optimization and Control

A Generalized Stability Theorem for Interconnections of Dissipative Behaviors

Dissipativity theory enables the stability analysis of complex, interconnected systems through the physical properties of their components. Celebrated examples of such stability results include the small gain and passivity theorems. In this paper, we prove a unifying stability theorem for dissipative linear time-invariant (LTI) systems. We use the behavioral approach to express storage functions and supply rates in terms of quadratic differential forms (QDFs). Our main result establishes conditions for the stability of an interconnection of two LTI behaviors that are both dissipative with respect to general, dynamic supply rates. As special cases, we recover existing results for static supply rates, such as those capturing finite $L_2$-gain and passivity, as well as dynamic ones capturing, for example, delta dissipativity. The key idea of the paper is to construct a Lyapunov function as a sum of storage functions and a coupling QDF, thereby allowing for a richer class of Lyapunov functions than existing works dealing with just sums of storage functions.
Read the full abstract

Quantum Physics

Low-Depth Initial-State Preparation for Ground-State Energy Estimation of Two-Dimensional Strongly Correlated Systems

Quantum algorithms for ground-state energy estimation require initial states with non-negligible fidelity to the ground state. Guided by classical simulations, we design shallow circuits for the two-dimensional half-filled Hubbard model by first preparing an approximate Heisenberg ground state and then applying charge-fluctuation gates derived directly from the Schrieffer-Wolff (SW) generator. For the Heisenberg model, we demonstrate effective parameter transfer from a $4\times4$ lattice to lattices up to $10\times10$ without further optimization. We obtain promising lower bounds on the ground-state fidelity using tensor-network simulations, variational Monte Carlo reference states, and estimates of the ground-state energy and singlet gap. Exact $4\times4$ calculations show that the SW gates substantially improve the Hubbard ground-state fidelity of the embedded Heisenberg state. The successful Heisenberg parameter transfer supports the use of these small-lattice results to design shallow Hubbard circuits on larger lattices beyond the reach of classical simulations. For a $10\times10$ lattice, the estimated preparation cost is on the order of $10^5$ $T$ gates, including Clifford+$T$ synthesis, which is well within the megaquop regime. These results offer a route to low-cost initial-state preparation for ground-starongly correlated systems.
Read the full abstract

Machine Learning

Sylvas: Synergistic Learning Value based Device Scheduling in Federated Continual Learning

Federated continual learning (FCL) enables shared global models to continuously adapt to distributed and non-stationary data streams, making it important for Internet of Things applications such as intelligent transportation, industrial monitoring, and unmanned systems. Under spatio-temporal data distribution dynamics and label scarcity, a key challenge is how to quantify the contribution of each edge device to global learning performance and schedule the most valuable devices under resource constraints for timely model updating. This article presents Sylvas, a synergistic learning value based device scheduling framework for FCL at the wireless edge. Sylvas evaluates the learning value of distributed data from two perspectives: distributional value, which characterizes the contribution of device data to global model learning from a spatio-temporal distribution perspective, and label value, which captures the quantity and reliability tradeoff of pseudo-labeled data. By integrating these factors into a synergistic learning value metric, Sylvas schedules devices with high learning value while satisfying communication and computation resource constraints. Case studies demonstrate that Sylvas supports timely model adaptation under spatio-temporal distribution dynamics and effectively exploits unlabeled data.
Read the full abstract

All authors Yuxuan Sun, Yuxuan Bai, Tan Chen, Sheng Zhou, Zhisheng Niu

Fluid Dynamics

Deep Koopman Sensing

Real-time reconstruction of fluid flows from sparse sensor measurements is important for both physical understanding and flow control. When first-principles models are too expensive for online data assimilation (DA), learned reduced-order models provide an efficient alternative, but are commonly optimized for forward prediction rather than state estimation. We propose Deep Koopman Sensing, a data-driven reduced-order DA framework that combines a nonlinear autoencoder with parameter-conditioned linear latent dynamics approximating the Koopman operator. We compare the proposed model with parametric dynamic mode decomposition (pDMD), a multilayer perceptron (MLP), and xLSTM across four benchmarks: 1D viscous Burgers, 2D flow past a cylinder, 2D dambreak, and 3D flow past a sphere. Our results reveal a marked distinction between forecasting and sensing: open-loop accuracy does not reliably predict assimilation performance, while Deep Koopman Sensing achieves the lowest assimilation error across all four benchmarks. More importantly, with an extended Kalman filter, incorporating sensor measurements improves the estimates of both linear latent models across all four benchmarks, whereas it degrades the nonlinear models, despite their strong open-loop performance. With ensemble filtering, the nonlinear models are no longer degraded by assimilation, while the Koopman model still attains the lowest assimilation error. These results show that latent dynamics should be designed for the downstream estimation task rather than selected solely for forecast accuracy, and demonstrate Koopman-based reduced-order modeling as an effective approach for real-time flow reconstruction from sparse, streaming measurements.
Read the full abstract

High Energy Astrophysical Phenomena

Modeling Gamma-Ray Bursts Using CRISP

Astrophysical neutrinos are a clear identifier of photohadronic interactions in sources of Ultra-High-Energy Cosmic Rays (UHECRs). As the Pierre Auger Observatory has shown that UHECRs exhibit a composition heavier than protons at Earth, we expect that these heavy primaries undergo nuclear cascades due to photodisintegration in the dense source environments from which they originate. In this contribution, we will derive the neutrino spectra including those originating from nuclei undergoing photohadronic interactions. This will be explored in the context of modeling the emission region of a gamma-ray burst. Photohadronic interactions will be modeled using the new framework for Cosmic Ray Stochastic Interactions for Propagation (CRISP), which employs an analytic approach to compute the underlying probabilistic description of UHECR interactions.
Read the full abstract

Audio and Speech Processing

Directivity-Conditioned Low-Latency Neural Filtering for Speech Enhancement in Hearing Aids

Latest advances in neural directional filtering show exceptional results in adapting the direction and shape of directivity patterns during the inference phase. However, in the existing methods for adapting directivity patterns during inference, important real-world constraints have been disregarded. Particularly for hearing devices, scenarios are often much more dynamic, microphone positions vary with head diameter and hearing aid placement, head-shadow effects occur, and strict latency constraints apply. In this work, we propose a novel low-latency (10 ms) deep neural network (DNN) taking the above requirements of hearing devices into account. As in recent work, we use feature-wise linear modulation (FiLM) to steer the directivity patterns during testing. To preserve the desired directivity pattern, a loss function is proposed that maintains the spectral cross-channel relationships. Interestingly, we are able to achieve similar results to methods with relaxed latency constraints.
Read the full abstract

All authors Lennart Uphaus, André Merboldt, Markus Hofbauer, Timo Gerkmann

Computation and Language

Enabling Streaming User Transcription in Full-Duplex Speech-to-Speech Models

Full-duplex speech-to-speech (S2S) models enable natural conversational AI by allowing simultaneous listening and speaking. However, these models typically lack inherent user speech transcription, which is essential for applications such as conversation logging, accessibility features, and quality monitoring. In this work, we propose an efficient method to add streaming ASR capabilities to an existing duplex S2S model by introducing a lightweight ASR head in parallel to the agent text head. Our approach requires minimal additional parameters and no significant architectural changes to the base S2S model, enabling real-time user transcription while preserving full-duplex conversational capabilities including turn-taking and barge-in handling. Experimental results demonstrate that our method achieves streaming average WER of 10.21% on the HuggingFace Open ASR Leaderboard within the duplex S2S framework. Additionally, we show that the same architecture trained as a standalone streaming ASR model achieves competitive results (7.73% WER) compared to current SOTA models. We will open-source our training and inference code to facilitate further research in joint streaming ASR and S2S modeling.
Read the full abstract

All authors Ke Hu, Nourchene Ferchichi, Edresson Casanova, Ankita Pasad, Elena Rastorgueva, Chen Chen, Nithin Rao Koluguri, Piotr Zelasko, Yifan Peng, Hainan Xu, Zhehuai Chen, Boris Ginsburg

Computation and Language

Sequential Adapter Stacking for Cross-Lingual Low-Resource ASR

Extending large-scale multilingual automatic speech recognition (ASR) models to low-resource languages remains challenging. Model performance is skewed toward high-resource languages and degrades sharply for languages with limited labeled data and pre-training exposure. To address this, we investigate parameter-efficient approaches for transferring knowledge from resource-rich source languages to low-resource target languages on Whisper. Alongside warm initialization and attention-based fusion, we propose Sequential Adapter Stacking, which places a trainable target-language adapter on top of a frozen source-language adapter. Under controlled experiments, these approaches are evaluated on three target languages unsupported by Whisper – Asturian, Assamese, and Xhosa – using source languages with varying degrees of relatedness. Sequential Adapter Stacking with the closest related source consistently and significantly outperforms full fine-tuning across the three targets, with 5–8% relative WER reductions. These gains largely persist with only one hour of target training data.
Read the full abstract

Quantum Physics

Optimal entanglement-assisted source coding under a balanced-difference promise

Entanglement can reduce the communication required for coding tasks, but establishing the minimum achievable cost is essential to understanding its limits. We address this question in a zero-error source-coding task where Alice receives a word and Bob knows an unordered pair of candidates containing it. Alice does not know the pair and must enable Bob to identify her word without error using shared entanglement and one classical message. The candidates satisfy a balanced-difference promise: for words in $\mathbb{Z}_q^n$ with $n=q\ell$, each residue modulo $q$ occurs exactly $\ell$ times in their coordinatewise difference. For all integers $q\geq2$ and $\ell\geq1$, we prove that the task requires exactly $n$ messages when $(q-1)\ell$ is even and two messages when it is odd. These minima allow arbitrary finite-dimensional shared states independent of the inputs and arbitrary local measurements. In even parity, this establishes optimality of an existing entanglement-assisted protocol. In odd parity, an explicit deterministic protocol achieves the optimum of one bit without entanglement. Our proof combines Fourier analysis with a combinatorial counting argument to determine the smallest eigenvalue of the associated graphs. In even parity, this resolves the spectral assertion of Cao et al.'s Conjecture 6.3 for balanced cyclic generalized Hadamard graphs. Together with an explicit odd-parity bipartition, this determines the quantum chromatic number as $n$ in even parity and $2$ in odd parity, where the classical chromatic number is also $2$. All lemmas, theorems, and corollaries are formalized and verified in Lean.
Read the full abstract

Earth and Planetary Astrophysics

Thermophysical characterization of the potentially hazardous near-Earth object 2024 YR4

The asteroid 2024 YR4 is a potentially hazardous Apollo-type near-Earth object. Its highly eccentric orbit produces close Earth-Moon encounters roughly every 4 years. A collection of light curves, including JWST-NIRCam data, allowed us to constrain the rotation period, spin pole, and convex shape solutions for the object. We analyzed two sets of JWST-MIRI observations obtained in three filters on 26 March 2025. The extracted photometry reveals rotationally driven thermal flux variations. The combination of a spin-shape model from light-curve inversion and the absolutely calibrated 3-band MIR light curves is well reproduced by a convex spin-shape solution with a spin pole at (l,b) = (232, -11), a sidereal period of 19.4633 min, and axis ratios of a/b ~1.28 and b/c ~1.45. The corresponding equivalent-volume diameter is 60.8 +1.8/-3.6 m (pV= 0.11 +0.05/-0.03, assuming HV = 24.14+/-0.25 mag). A high thermal inertia and low surface roughness are required to explain the observed MIRI data. The near-Earth asteroid thermal model has limited reliability for this object, requiring an extreme beaming parameter >3, whereas the fast-rotating model and a full thermophysical model provide consistent and reliable estimates of the size and albedo. The combination of high eccentricity and high thermal inertia implies seasonal skin depths of up to ~3 m and radiative timescales of weeks to months. Although seasonal heating is negligible for the March 2025 observations because of the near-equator-on viewing geometry, the seasonal Yarkovsky drift is expected to be dominant over the diurnal component and might become measurable during future encounters. Although JWST astrometry from February 2026 rules out any possible impact on the Earth or the Moon for at least 100 years, the spin, shape, and radiometric constraints presented here provide an important benchmark for future potentially hazardous objects.
Read the full abstract

All authors T. G. Müller, E. M. MacLennan, A. Y. Burdanov, B. J. Holler, A. S. Rivkin, J. de Wit, P. Pravec, M. Micheli, K. Muinonen, D. Farnocchia

Systems and Control

Extended Version: Storage-Based Strategic Manipulation of Constraint-Binding Patterns in Power Networks

This paper studies the strategic market participation of a monopolistic energy storage aggregator (ESA) in a day-ahead electricity market. The ESA coordinates geographically distributed storage units, submits a coordinated bid for its portfolio, and may hold financial transmission rights (FTRs). The system operator clears the market through a network-constrained, multi-period economic dispatch, determining generation and load schedules, nodal prices, energy-market payments, and FTR payoffs. We formulate the ESA–system-operator interaction as a Stackelberg game and characterize its equilibrium through a constraint-binding-pattern decomposition of the market-clearing problem. Beyond enabling equilibrium computation, the framework reveals how the ESA can increase its profit by strategically inducing or avoiding particular constraint-binding patterns. It also establishes a novel welfare result: although strategic storage without FTRs is known to weakly improve social welfare relative to the no-storage case, certain FTR positions can overturn this guarantee by strengthening the ESA's incentive to induce particular patterns, causing social welfare to fall below the no-storage level. Motivated by these findings, we develop two system-operator mechanisms for limiting undesirable ESA behavior and its adverse effects on market outcomes and social welfare. Finally, a three-bus study illustrates the theoretical findings, while IEEE test systems demonstrate the scalability of the proposed method.
Read the full abstract

Combinatorics

Direct Generation of a Somos-4 Sequence from an Algebraic Generating Function

We construct a five-parameter quadratic algebraic generating function whose coefficient sequence has prescribed initial Hankel determinants $(\lambda,m,r,\eta)$ and whose Hankel transform belongs to the Somos-4 family $A(1,\tau)$. The construction starts from a Stieltjes continued fraction whose coefficients are generated by an alternating recurrence compatible with the Somos-4 relation. We derive an explicit quadratic equation for the resulting generating function and give the corresponding coefficient recurrence. A subtle feature of the construction is the nonuniqueness of a coefficient sequence determined solely by its ordinary Hankel transform. To select a canonical representative, we additionally prescribe the shifted Hankel determinants obtained from the same Somos-4 orbit advanced by two indices. This companion condition determines the odd and even Stieltjes coefficients separately and removes the remaining freedom in the continued-fraction representation. We also analyze the two algebraic branches at the origin, where they coalesce, and derive a desingularized recurrence for the coefficients. The resulting construction provides a direct algebraic generating-function realization of a general five-parameter family of Somos-4 Hankel transforms.
Read the full abstract

Statistics Theory

Covariance and Principal Component Analysis on Riemannian Manifolds and Graphs

We develop notions of covariance and principal component analysis (PCA) for probability measures on Riemannian manifolds of bounded geometry and a discrete counterpart for distributions on the vertex sets of weighted simple graphs. Rather than anchoring variation at the Fréchet mean, whose uniqueness and usefulness can fail in this context, we consider variation about every point. This leads to a representation of each point on the manifold by a vector field derived from the heat kernel and thus to a map of the underlying manifold into a Hilbert space of vector fields, in which covariance and PCA carry over from the Euclidean setting. A distinguishing feature of this formulation is that one typically obtains infinitely many principal components, capable of capturing highly nonlinear geometric features and patterns of variation. We prove an embedding theorem for the mapping into the space of vector fields and develop two computational reductions of the theory: RiePCA, a finite-dimensional reduction based on vector fields supported on finitely many points, and GraphPCA, a discrete formulation for measures on weighted graphs in which vector fields assign orientations and magnitudes to edges. Numerical examples and experiments illustrate the behavior of the proposed methods on both synthetic and real data.
Read the full abstract

Optics

Tailoring complex coupling for efficient synchronization of semiconductor lasers

A grand challenge for laser synchronization is the intrinsic frequency disorder. We build an experimental platform for programming the complex coupling coefficients of semiconductor lasers to promote frequency and phase locking. The coupling amplitude between every pair of lasers in an array is tuned to scale with their frequency difference. Coupling phases are scanned to meet the phase matching conditions for constructive interference of coupled fields. The minimum coupling budget for global synchronization is significantly reduced, as shown experimentally for three vertical-cavity surface-emitting lasers with intrinsic frequency detuning. The customized coupling offers an efficient approach to stable synchronization of semiconductor laser arrays for coherent beam combining.
Read the full abstract

All authors Nathan Vigne, Amit Pando, Li-Li Ye, Mehmet Bütün, KyeoReh Lee, Ying-Cheng Lai, Nir Davidson, Fan-Yi Lin, Hui Cao

Mesoscale and Nanoscale Physics

Effects of residual exchange coupling on simultaneously driven spin qubits

Exchange coupling and microwave drives are widely used control mechanisms for spin qubits. However, the influence of exchange coupling on microwave-driven spin dynamics is not fully understood. We report simultaneous drive measurements of two spin qubits as a function of exchange coupling and drive power. When the Rabi frequency exceeds exchange, we observe a beating pattern in the Rabi oscillations. In the opposite limit, two exchange-split patterns emerge in the Rabi chevrons. We find that these exchange-induced effects are suppressed when the difference in the Rabi frequencies exceeds the exchange coupling. A theoretical model is developed that reproduces the main features in our data.
Read the full abstract

All authors Heun Mo Yoo, Tanner M. Janda, Victor Yu, Michael J. Gullans, Adam R. Mills, Jason R. Petta

Cosmology and Nongalactic Astrophysics

Toward Precision Kinetic Sunyaev-Zel'dovich Cosmology—1. The Matter - Momentum Bispectrum at One-Loop Order

The kinetic Sunyaev-Zeldovich (kSZ) effect is a sensitive probe of the free-electron momentum field, tapping into spatial correlations relevant for both large-scale cosmology and smaller-scale astrophysics. Precision kSZ measurements will provide a robust channel for constraining inflationary physics and equivalence principle violation, while simultaneously characterizing the small-scale gas distribution and associated baryonic feedback. However, robust kSZ measurement for cosmological applications is challenging, as the primary source of large-scale information is convolved with unknown small-scale dynamics in a degenerate combination. In this paper we address this challenge using the Effective Field Theory (EFT) framework that provides a principled way of treating the effects of small-scale dynamics on large scale-structure. Extracting the kSZ signal can be reformulated as a measurement of the one-loop galaxy-galaxy-electron momentum bispectrum. We compute here a pure dark matter version of this bispectrum in EFT and compare it to simulations. We derive the EFT counterterms that capture small-scale backreaction and find that two of them acquire a transverse direction. One transverse counterterm gets generated entirely by the vorticity of the velocity field. Our computation of the bispectrum dipole agrees with simulation data to 5% up to k_max = 0.23 h/Mpc, which is significantly larger than the reach of tree level theory, which breaks down beyond k_max = 0.07 h/Mpc (at redshift z = 0.5). For the bispectrum computed with the projected momentum field, relevant to measuring the kSZ field, we find a similar k_max, and make the first detection of the transverse counterterms from simulations. Our results suggest that one-loop bispectrum computations can play an important role in advancing kSZ cosmology.
Read the full abstract

Image and Video Processing

OASIS: Online Adaptive Video Compression via Closed-loop Feedback Control

Computer vision systems are a key building block in an autonomous vehicle, responsible for a range of perception tasks. However, they incur massive data transmission over long communication links from multiple cameras, creating a critical bandwidth and energy bottleneck. Although conventional codecs such as H.264 can reduce data rates, they are ill-suited for real-time vision systems due to high processing latency and energy consumption, as well as their reliance on static user-defined compression settings. In light of these challenges, we propose OASIS, an adaptive video compression framework that integrates lightweight in-sensor compression with task-aware compression ratio control. Based on the real-time task performance, it dynamically updates the optimal compression ratio. Experimental results demonstrate that OASIS generalizes across multiple vision tasks, achieving on average a 6x data compression, 5.8x reduction in link power consumption, and 2.5x reduction in link latency, with at most 1.5% performance degradation.
Read the full abstract

Cosmology and Nongalactic Astrophysics

Using the Circumgalactic Medium of Dwarf Galaxies as a Calorimetric Dark Matter Detector

Many well-motivated dark matter (DM) scenarios, including dark photon DM, axion-like particles, sterile neutrinos, and primordial black holes, can deposit energy into ordinary matter and significantly raise the thermal energy of astrophysical gas. We present a simple, model-independent implementation of DM-induced heating in the RAMSES hydrodynamic simulation code, parameterized by an effective local heat-transfer rate $\Gamma_{\rm heat}$. Using simulations of an isolated gas-rich dwarf galaxy with properties similar to WLM, we find that DM heating substantially alters the thermal structure of the circumgalactic medium (CGM) while leaving the star formation rate and the dense ISM nearly unchanged. For $\Gamma_{\rm heat} = 10^{-27}$ to $10^{-25}\ {\rm s}^{-1}$, the neutral hydrogen column at 10 kpc is suppressed by 0.6 to 3.9 dex relative to the control simulation. Similarly, the C II and Si II column densities decrease by 4.1 and 5.5 dex, and by 2.2 and 3.0 dex for C IV and Si IV. These signatures are accessible to quasar absorption-line spectroscopy, so CGM observations of nearby field dwarf galaxies provide a promising avenue for constraining dark matter models that inject heat into ordinary matter.
Read the full abstract

Neurons and Cognition

When Teachers Smile or Frown: A Profile-Based Analysis of Achievement Emotions

Achievement emotions shape how students engage with and learn from academic tasks, yet most studies examine individual emotions rather than co-occurring affective profiles and their dynamics. We examined latent achievement-emotion profiles and their transitions following exposure to different instructor facial expressions during a video lecture. Self-reported data from 78 Grade VII and VIII students revealed three profiles: enthusiastic, demotivated, and vulnerable. Profile transitions differed across instructor conditions, with happy expressions favouring more adaptive transitions and angry expressions favouring transitions toward demotivation. Exploratory factor analysis and Bayesian structural modelling further identified preparedness and cognitive restraint as regulatory dimensions associated with profile switching.
Read the full abstract

Graphics

TopoRig: Topology-Agnostic Facial Rigging via Multi-Source Supervision

Automatic facial rigging across heterogeneous mesh topologies remains challenging because high-quality expression supervision is often tied to canonical templates, while deformation transfer to arbitrary meshes can introduce geometric artifacts and correspondence errors. We present TopoRig, a topology-agnostic facial rigging framework that predicts FACS-conditioned deformations directly on input mesh vertices while preserving the original topology. Starting from the ICT FaceKit expression model, we construct complementary supervision from accurate but template-biased common-topology rigs, topology-diverse but noisier transferred rigs, and targeted image-based cues for controls poorly captured by geometric transfer. TopoRig combines local surface geometry, landmark-relative semantic features, global shape context, and FACS controls to predict per-vertex displacements. We train on 3,496 generated identities using 45 non-gaze expression controls from the 53-control ICT FaceKit vocabulary. On held-out identities and unseen mesh topologies, TopoRig more faithfully reproduces the reference expression space than prior neural facial-rigging methods, while qualitative results show consistent localized deformations across diverse character geometries. Ablations demonstrate that semantic landmark features and complementary supervision improve cross-identity and cross-topology generalization. Overall, TopoRig amortizes heterogeneous and imperfect expression supervision into a single topology-preserving deformation model.
Read the full abstract

All authors Andrew Fleet, Soroush Mehraban, Vida Adeli, Cole Clifford, Babak Taati

Computation and Language

Look Before You Leap: Factual Decoding with Internal Attribution Signals

Hallucination remains a critical challenge in large language models (LLMs), where early factual errors compound through autoregressive generation in a snowballing effect that neither post-hoc correction nor weight-level intervention can effectively preempt. We propose DescaPE (DEcoding Signal Control Against Path Error-snowballing), a decoding framework that leverages internal model signals to suppress hallucination-prone trajectories at inference time. Through sliding-window MLP ablation, we identify a factual-salient layer span within LLMs whose derived signal is selectively elevated for factual tokens and exhibits anomalous spikes at hallucination-prone steps. We train a lightweight probe to approximate this signal from a single forward pass and integrate it into candidate scoring to penalize high-risk continuations while rewarding factually grounded ones. Experiments across five factuality benchmarks on three LLMs demonstrate that DescaPE achieves factuality improvements over decoding-time baselines in multiple settings, while incurring only 1.10x latency overhead in our efficiency evaluation. Our code is available at https://github.com/hayeonggg/DESCAPE.
Read the full abstract

All authors Hayeong Ryu, JungMin Yun, Byeonggeuk Lim, Sunhee Jo, YoungBin Kim

Artificial Intelligence

Design of a Deep Learning Credit Risk Early Warning System Integrating Multi-source Heterogeneous Data

Advancements in data fusion and real-time analytics technologies have opened new avenues for addressing complex domain challenges. Financial risk early warning systems often suffer from inefficiency due to information silos and monitoring delays. This paper proposes a credit risk early warning system based on heterogeneous information fusion. The system employs a model architecture integrating deep neural networks and attention mechanisms to extract multidimensional features from diverse data sources such as transaction behaviors and social networks, thereby establishing an early identification mechanism for corporate and individual credit risks. System testing demonstrates that this approach significantly enhances the accuracy and timeliness of risk warnings, outperforming traditional rule-based engine solutions. The findings offer innovative insights for early intervention in financial risks, holding practical significance for safeguarding financial stability.
Read the full abstract

All authors LiYang Wang, Zhen Zhong, Zhen Tian, Keyu Chen Keyu Chen

Computation and Language

Merging the Knowledge of LLMs for Automatic Speech Recognition

Automatic speech recognition (ASR) systems, trained on paired speech-text data, have been improved by leveraging language models (LMs) trained on text-only data. LM fusion methods such as shallow fusion and density ratio are well-established methods that incorporate external LMs during ASR decoding. However, they incur additional computational costs due to LM inference, which is particularly problematic for recent larger LMs. In this study, we propose incorporating external LMs via model merging. This method integrates the LMs directly into the parameters of an LLM-based ASR model, requiring no additional computational cost at inference. We formulate domain extension and transfer via arithmetic operations on LoRA parameters. Experimental evaluations were conducted for the domain adaptation of LLM-based ASR trained on CSJ and LibriSpeech. We show that our LM merging consistently improved the ASR performance in the target domains, without degrading inference speed or memory footprint.
Read the full abstract

High Energy Physics (Phenomenology)

Explaining the LZ High-Energy Recoil Event with Inelastic Sneutrino Dark Matter in Supersymmetry

We study an inelastic sneutrino dark matter origin of the LZ high-energy event in the inverse-seesaw next-to-minimal supersymmetric standard model. Approximate lepton-number conservation protects the small mass splitting, while singlet interactions separate freeze-out from solar-neutrino production. The LZ recoil and solar constraints overlap even for a fully cooled population in equilibrium. Three thermal mechanisms reproduce the relic abundance and satisfy all relevant constraints. Photon-dominated annihilation produces gamma-ray boxes, offering a complementary test of this interpretation.
Read the full abstract

Risk Management

Quantifying the 2027 Solvency II Risk Margin Reform

The 2027 Solvency II reform recalibrates the Risk Margin by reducing the prescribed cost-of-capital rate from 6% to 4.75% and introducing a time-dependent attenuation of future Solvency Capital Requirements. This paper develops an analytical and numerical framework for characterizing the effect of the final regulatory calibration. By normalizing discounted projected capital requirements into a probability distribution over run-off time, we obtain an exact representation of the ratio between the revised and previous Risk Margins. The framework yields sharp universal and horizon-specific bounds, characterizes the effect of later capital timing through stochastic dominance, and shows that mean run-off time alone does not determine the reform effect when temporal dispersion varies. Additional bounds are derived conditional on mean run-off time and horizon. Conditional on a given projected SCR path, the mechanical reduction lies between 20.83% and 60.42%. For proportional Best Estimate projections, an exact covariance decomposition identifies how departures from proportionality affect the relative reform ratio. Holding the projected SCR path fixed, the revised formula has a lower direct interest-rate semi-elasticity than the previous calibration. A reduced-form stochastic extension further quantifies convexity effects arising from uncertainty in capital persistence. Numerical applications reconstruct Risk Margin calculations from published actuarial run-off profiles and complement them with controlled long-horizon, interest-rate, and persistence experiments. The results show that the reform effect is governed by the temporal structure of future capital and provide tractable tools for assessing projected capital profiles, Risk Margin simplifications, and direct discount-rate sensitivity.
Read the full abstract

Machine Learning

A Language-Guided Multimodal Foundation Model for Zero-Shot and Multi-Task Brain Signal Analysis

Brain signal analysis is essential for both neuroscience research and clinical diagnostics, yet current approaches face critical limitations. End-to-end models require task-specific retraining and exhibit limited generalization, while pre-trained models lack semantic depth and still depend on extensive fine-tuning. Meanwhile, general-purpose multimodal foundation models, though powerful in other domains, struggle to interpret brain signals due to representational misalignment and lack of domain knowledge. This study introduces a multimodal foundation model for zero-shot and multi-task brain signal analysis (METIS) through a unified language-signal alignment framework. METIS is pretrained on the largest and most diverse brain-signal corpus to date, comprising over 70,000 h of recordings from more than 11,000 subjects across 20 datasets. In a comprehensive zero-shot evaluation across 12 datasets, METIS outperformed the leading generalist model by over 20.9% in average accuracy. Remarkably, without any fine-tuning, METIS's performance matches or exceeds that of supervised, task-specific models. Furthermore, METIS demonstrates exceptional data efficiency and strong generalization, achieving an average AUROC advantage of over 16.0% in few-shot settings and 15.9% in cross-dataset transfer. This work establishes a new paradigm for general-purpose brain signal analysis, paving the way for next-generation neurotechnology.
Read the full abstract

All authors Mingzhi Chen, Yiyu Gui, Guibo Luo, Yuchao Yang

Quantum Physics

Pauli spectrum and nonstabilizerness of random fermionic Gaussian states

We characterize the magic of random fermionic Gaussian states through their Pauli spectrum and their stabilizer entropies. For the Haar ensemble of Majorana Gaussian states we derive closed finite-$N$ expressions for the average stabilizer purities and reconstruct the Pauli spectrum exactly, as a mixture of products of beta-distributed variables resolved by Majorana weight. The stabilizer entropies, and their filtered versions, freeze at Rényi index $q_c=2$: for $q>2$ the magic density is $1/(q-1)$, controlled by rare weight-two Majorana strings. Resolving these contributions requires exponentially many characteristic samples, making direct sampling estimates of the stabilizer entropies inefficient in this frozen regime. For Haar-random Gaussian states at fixed filling, we express the averaged stabilizer purities at positive integer $q$ as coefficient integrals over $\mathrm{SU}(2q)$, with dimension independent of $N$, and evaluate the $q=2$ case exactly. We establish that the frozen density describes typical states in the Majorana and half-filled number-conserving ensembles and relate these ensembles to the corresponding $\mathrm{SYK}_2$ ground states. Finally, we show that typical fermionic Gaussian states retain certified magic after discarding any fixed fraction $\kappa<\kappa_\star^{\rm G}\simeq0.7613$ of their modes in the thermodynamic limit, remaining magical even beyond the $2/3$ threshold for Haar-random states in the full Hilbert space.
Read the full abstract