Peer Revenue Publish or profit

Papers

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

All of arXiv

Monday 7 September

Computation and Language

A*-Thought-V2: Efficient Latent Reasoning via Geometric Dynamics of LLM

Chain-of-Thought (CoT) improves the reasoning ability of Large Language Models (LLMs) but incurs substantial computation and context costs. Existing methods either lose intermediate information through hard pruning or lack a principled criterion for continuous compression. We present A*-Thought-V2, a geometric dynamics of LLM guided framework that models CoT as a hidden-state trajectory and replaces hard deletion with an explicit-implicit interleaved latent architecture. After projecting question, step, and solution representations into a 3D PCA space, it measures alignment between each local transition and global question-to-solution direction. Aligned steps remain explicit text, whereas deviating steps are compressed into continuous latent tokens. Directional angles capture both local semantics and reasoning dynamics: small angles indicate direct execution and answer formation, while large angles more frequently involve checking, correction, and branch exploration; their temporal variation reveals exploration, convergence, and refinement stages. To train this architecture, we introduce stepwise embedding forcing, which pools each redundant step into a single latent embedding, and label forcing, which supervises that latent token with a soft multi-modal vocabulary distribution instead of a hard one-hot label. Experiments on Qwen3.5-9B and Qwen3.6-27B across six in-domain and out-of-domain benchmarks show that A*-Thought-V2 improves average accuracy by up to 2.6% while reducing response length by up to half, increasing Accuracy per Computation Unit by 2.29$\times$, and reducing preprocessing and training time by 94.6% and up to 80.3%, respectively. Representation analyses suggest that latent states form a compact region distinct from textual states, while higher entropy at latent-token positions reflects broader soft targets that encourage richer step-level feature learning.
Read the full abstract

All authors Xiaoang Xu, Siyuan Liu, Shuo Wang, Junlan Feng, Fanyu Meng, Zhu Zhang, Jixun Wang, Xiaorong Wang, Zihan Zhou, Xin Li, Chaojun Xiao, Yiming Zhang, Huijia Wu, Liuyu Xiang, Peipei Li, Zhaofeng He

Mesoscale and Nanoscale Physics

Creation and characterization of leviton excitations in tight-binding chains

Levitons are minimal-excitation electronic wave packets generated by Lorentzian voltage pulses and constitute a central resource for electron quantum optics. Their creation and defining properties are usually formulated in continuum scattering descriptions, whereas many candidate platforms for integrated electronic quantum circuits are finite, discrete, and strongly shaped by lattice dispersion. We study leviton generation in finite one-dimensional tight-binding chains of non-interacting fermions driven by time-dependent voltage pulses. Using the single-particle density matrix, we resolve the excitation above the initial Fermi sea and quantify its quality through the average excitation number and its fluctuations. We find that clean leviton-like states emerge only in an intermediate regime where the pulse is slow enough to be resolved by the dynamics on the lattice, but not so slow that truncation and finite-size effects distort the Lorentzian profile in the time domain. Lorentzian pulses (even if truncated) systematically outperform non-Lorentzian pulse shapes in approaching the low-noise limit with increasing system size. We further identify finite-lattice signatures associated with band filling, pulse amplitude, linear voltage-drop geometry, and residual deviations from the continuum integer-charge condition. These results establish a microscopic framework for understanding leviton formation beyond the ideal continuum limit and for evaluating lattice-based platforms for coherent few-electron transport.
Read the full abstract

Probability

Markov Chain CLTs: Resolving Open Problems

Markov chain central limit theorems (CLTs) and their associated variances are very important for implementing Markov chain Monte Carlo algorithms among other applications. Häggström and Rosenthal (2007) presented various results regarding the equality of different formulae for this variance and also posed seven open problems. We resolve all seven in this paper. Our main result is a general characterization of Markov chain CLTs: for stationary and ergodic Markov chains, we prove that the normalized partial sums satisfy a $\sqrt{n}$-CLT if and only if the Poisson equation admits a finite solution with square-integrable martingale differences whose second moment equals the CLT variance. In particular, this result does not require reversibility of the chain or square integrability of the function. It implies that Roberts' holding-probability condition always precludes a $\sqrt{n}$-CLT, and for reversible chains, whenever a $\sqrt{n}$-CLT holds, the variance limit is finite if the function is square-integrable, otherwise undefined. In addition, we develop a general principle that expresses Fourier coefficients as the autocovariances of an ergodic nonreversible Markov chain.
Read the full abstract

Earth and Planetary Astrophysics

Sea Ice as an Origin of Life Location for Hycean and Ocean Worlds

Water worlds with potential deep oceans are at the forefront of currently observable exoplanetary habitability. The standard requirements for habitability are the presence of an energy source, a source of nutrients, and a solvent, all of which can be met in exoplanets with deep liquid water oceans. Such planets may be numerous, ranging from hycean worlds and ocean worlds to terrestrial-size planets with enhanced water content. However, abiogenesis requires additional conditions, namely a mechanism to accumulate nutrients to the high concentrations required for prebiotic chemistry, and an energy source to drive the chemistry. A deep ocean, and in some cases the high-pressure ice layers underneath, has traditionally been thought to prevent the required concentrations. Here we outline a location where nutrients could be concentrated in sufficiently high quantities to make abiogenesis possible – at the interface between the open ocean and the surface ice shelf. Using a 3D General Circulation Model, we explore cases corresponding to several candidate habitable exoplanets which might have dayside sea ice. We discuss how the planet's climate influences the nature and behaviour of the sea ice. We investigate the ability of small-scale ice formation and melting processes to concentrate nutrients to prebiotically useful levels, and also explore the potential for meteoritic impactor fragments to provide a platform for prebiotic chemistry. We find that both of these mechanisms may plausibly provide concentrated chemical feedstock for prebiotic chemistry, which adds an important new perspective when evaluating the prospects of habitability in planets with deep oceans.
Read the full abstract

Combinatorics

Unbalanced distance-biregular graphs with girth deficiency two

Let $\Gamma=(V,E)$ be a distance-biregular graph with even diameter $d$ and girth $2d-2$. On the edge-set $E$ we define relations that form an association scheme; this is done by considering the linear span of the corresponding adjacency matrices, and, by means of intersection diagrams, we prove that this is an algebra. Furthermore, we describe all the irreducible representations of this algebra, and, by considering the multiplicities of the 2-dimensional ones, we show that there are no such distance-biregular graphs with $28\le d\le 46$ and $d\ge 52$.
Read the full abstract

Machine Learning

Kalman Delta Networks: Uncertainty-aware Associative Memory

Linear attention enables efficient long-context inference by compressing token history into a fixed-size recurrent memory. This compression makes each update a trade-off between incorporating new information and preserving useful associations. Models such as DeltaNet, Gated DeltaNet, and KDA predict write strength from the current token representation, without explicitly tracking uncertainty in the stored memory. Yet this uncertainty matters: a new observation should have greater influence when the existing association is uncertain and less when it is already well supported. We introduce Kalman Delta Networks (KDNs), a family of linear-attention models that explicitly track memory uncertainty to guide each update. By formulating associative memory as a linear-Gaussian state-space model, KDNs propagate both the memory estimate and its uncertainty, using the Kalman gain to balance accumulated evidence against the reliability of new observations. This formulation also recovers standard delta-rule updates by replacing tracked covariance with a token-predicted isotropic surrogate. To support hardware-efficient training and inference, we derive Diagonal KDN and Isotropic KDN, which retain one uncertainty value per key channel and per head, respectively. Their uncertainty updates admit associative scans with logarithmic parallel depth, requiring only $O(d_k)$ and $O(1)$ auxiliary state per head. Across controlled pretraining at 750M and 1.3B parameters, both variants consistently improve perplexity and mean downstream accuracy over the evaluated state-of-the-art linear-attention baselines.
Read the full abstract

Computer Vision

VoT: Vision-of-Thought for Unified Multimodal Representation Alignment

Current text-to-image systems typically employ a "text encoder plus diffusion decoder" paradigm, in which text semantics directly modulate continuous latent noise. Despite their success, these methods lack an explicit, interpretable intermediate representation that effectively bridges high-level linguistic semantics and low-level visual signals. In this paper, we propose Vision-of-Thought (VoT), a framework that introduces a discrete visual-thinking layer between vision-language models (VLMs) and diffusion transformers (DiTs). Instead of treating VLMs merely as text encoders, we use them as multimodal planners that generate discrete VoT tokens representing high-level visual plans, such as objects and layouts, before rendering pixels. We train a specialized VoT tokenizer in the VLM semantic space with a closed-loop objective that combines VLM alignment, feature reconstruction, and vector-quantization losses. These objectives make the tokens semantically readable by the VLM while preserving the visual information needed for generation. Experimental results demonstrate that VoT improves semantic alignment and provides a structured interface for interpretable and controllable generation.
Read the full abstract

All authors Jingxiang Sun, Chao Liao, Zhengxiong Luo, Chaorui Deng, Chen-lin Zhang, Junke Wang, Ceyuan Yang, Haoqi Fan, Weilin Huang

Machine Learning

Latent-MoE: Domain-Aware Mixture-of-Experts for PDEs with Multi-Regime Physics

Physics-informed neural networks (PINNs) struggle on PDEs whose governing physics varies across the domain. We trace this to a structural property of standard coordinate networks: their neural tangent kernel (NTK) is translation-variant and lets training points of large coordinate magnitude disproportionately influence predictions elsewhere, producing long-range coupling and gradient conflict during training. We show analytically and empirically that mixture-of-experts (MoE) architectures with centered, compact-support routers yield a uniformly banded NTK whose kernel-regression weights decay exponentially with distance, localizing the learning. Building on this, we propose Latent-MoE, which interleaves domain-aware MoE blocks within a shared backbone. Unlike FB-PINNs or X-PINNs, which rigidly partition both the domain and the parameters so that the parameters on different subdomains are updated independently, Latent-MoE is designed to preserve the localization benefit of domain-aware routing while allowing capacity to flow across regions through the shared backbone. On standard homogeneous-physics benchmarks Latent-MoE is competitive with established baselines; on benchmarks with multi-stage time-variable physics, where global models and rigid domain decompositions both fall into spurious solutions, it improves over them by more than an order of magnitude, with markedly reduced gradient conflict during training.
Read the full abstract

Analysis of PDEs

Nonlinear Schrödinger systems with all attractive forces

In this paper we investigate in a systematic way the solution structure of nonnegative solutions for coupled nonlinear Schrödinger systems in the all attractive regime. For all frequencies equal case we obtain results on Morse index of synchronized positive vector solutions and nonnegative semi-vector solutions as well as their kernel of linearized systems at these solutions. We establish a variational characterization of synchronized positive vector solutions and as applications we give a new existence result about positive vector solutions for the general systems with arbitrary frequencies. We also examine the synchronization phenomenon of positive vector solutions, and prove that if the interaction matrix has exactly one positive eigenvalue, any positive vector solution is synchronized. Finally we provide examples of domains for which synchronization of positive solutions fails to hold if the interaction matrix has at least two positive eigenvalues. Our results reveal the effect of the spectral information of the interaction matrix of the couplings and geometry of a domain on the solution structure.
Read the full abstract

Probability

The Fenchel Game of Underdamped Langevin Dynamics: Insights into Accelerated Convergence

For $(Q_t,P_t)$ governed by suitably damped underdamped Langevin dynamics, we quantify the convergence in KL divergence of the positional marginal to a $\sigma$-strongly log-concave target $\pi(dq) = \frac{1}{Z}e^{-V(q)}dq$ as \begin{align*} \operatorname{KL}(\operatorname{Law}(Q_t) \| \pi) \leq e^{-\sqrt{\sigma} t}\operatorname{KL}(\operatorname{Law}(Q_{0}, P_{0})\, \|\, \Pi_0), \end{align*} where $\Pi_0$ denotes an appropriately selected reference measure. When $\pi$ is merely log-concave, the estimate \begin{align*} \operatorname{KL}(\operatorname{Law}(Q_t) \| \pi) \leq \frac{\tau^2}{t^2}\operatorname{KL}(\operatorname{Law}(Q_{\tau}, P_{\tau})\, \|\, \Pi_{\tau}) \end{align*} is derived, where $\Pi_\tau$ denotes another correspondingly chosen reference measure at time $\tau > 0$. Both rates match precisely the canonical rates of the corresponding accelerated gradient flows in $\mathbb{R}^d$. They are achieved by the novel interpretation of the underdamped Langevin dynamics as a combination of strategies in an online sampling game and by estimating the KL divergence using a cost function informed by fictitious competitors.
Read the full abstract

High Energy Physics (Phenomenology)

Higgsino Dark Matter Interpretation of the LZ High-Recoil Event in the GNMSSM with TeV-Scale Gauginos

The nuclear recoil event at approximately 248 keV reported by the LUX-ZEPLIN collaboration motivates an investigation of endothermic dark matter scattering. We study this within the General Next-to-Minimal Supersymmetric Standard Model (GNMSSM), with Higgsino-dominated neutralino DM undergoing the $Z$-mediated transition $\widetilde{\chi}_1^0N\to\widetilde{\chi}_2^0N$. In the conventional thermal Higgsino limit of the MSSM, the observed relic abundance selects a mass near 1.1 TeV, while a neutralino splitting of a few hundred keV typically requires gaugino masses of order $10^7$ GeV. In the GNMSSM, Higgsino-Singlino mixing introduces an additional contribution to the splitting that can cancel the gaugino-induced contribution, allowing sub-MeV splitting with multi-TeV gauginos. This mixing also modifies the inelastic scattering coupling and annihilation rates, while coannihilation with sleptons provides freedom in obtaining the observed relic abundance. We present six benchmark points with DM masses of 0.6-1.1 TeV, neutralino splittings of 333-350 keV, and gaugino masses of 2-5 TeV. These points reproduce the observed relic abundance and satisfy direct-detection, Higgs, flavor, and collider constraints. Within the Standard Halo Model and extended-likelihood analysis, all six points yield $\Delta\chi^2<1$ relative to the best fit. Our results show how the GNMSSM can accommodate the LZ high-recoil event without an ultraheavy gaugino sector. A quantitative assessment of solar-capture and neutrino-telescope constraints remains necessary for establishing viability.
Read the full abstract

Mathematical Physics

Equation of Motion for Thermodynamic Systems and Their Dynamical Equivalence with Mechanical and Electrical Systems

Changes in the internal energy of a thermodynamic system are fundamentally described by the first law of thermodynamics. However, the first law has a highly compact form that does not explicitly reveal the dynamics of the system or provide a corresponding equation of motion. While equations of motion can be written explicitly for mechanical and electrical systems, an analogous dynamical formulation for thermodynamic systems has, to our knowledge, not been established directly from the first law. In this work, we show that, through dimensional reduction, the first law can be transformed from its compact thermodynamic form into an explicit equation of motion for an open thermodynamic system. Remarkably, the resulting equation has the same mathematical structure as the equations governing a driven damped harmonic oscillator and a driven series RLC circuit. Detailed dimensional analyses of the derived equation and the effective coefficients further confirm the dimensional consistency of the formulation. We further perform a systematic comparison of the mechanical, electrical, and thermodynamic systems and show that, despite their distinct physical origins, they share the same universal dynamical structure. Finally, we construct a Lagrangian formulation of the thermodynamic system and establish its correspondence with the Lagrangian descriptions of the mechanical and electrical systems. These results demonstrate that an explicit equation of motion can be generated directly from the first law of thermodynamics and reveal a common dynamical framework underlying thermodynamic, mechanical, and electrical systems.
Read the full abstract

Optimization and Control

Configuration-Constrained Tube MPC for Difference-of-Convex Nonlinear Systems

This paper develops a convex tube model predictive control formulation for constrained nonlinear systems. We consider dynamics described by a discrete-time state-space model with parametric and additive uncertainty that admits a difference-of-convex decomposition. Convex directional bounds of the nonlinear dynamics are combined with configuration-constrained polytopic tubes, whose predefined combinatorial structure yields an affine parameterization of their vertices. The resulting finite-dimensional sufficient conditions certify robust one-step tube propagation under parametric uncertainty and additive disturbances, while allowing the tube geometry and an associated vertex control law to be optimized jointly in a single convex program. An implicit terminal condition guarantees recursive feasibility and convergence of the predicted tube to a target robust control invariant set. Numerical results illustrate the closed-loop properties and the trade-off between geometric flexibility and computational complexity.
Read the full abstract

All authors Filippo Badalamenti, Jose A. Borja-Conde, Filiberto Fele, Teodoro Alamo, Daniel Limon

Computation and Language

You Can't Prefer Emotions You Don't Sample: Intensity Undershoot in DPO-Tuned LLMs

Ask a language model to respond "very excitedly," and its output is typically only mildly more energetic. We quantify this effect. We condition an instruction-tuned LLM on a continuous Valence-Arousal (VA) target, where valence measures how pleasant a state is and arousal how activated it is, measure the achieved affect with a frozen regressor, and sweep the requested target from -1 to +1. The response moves far less than asked: the gain, the slope of achieved against requested affect, is only 0.26 for valence and 0.13 for arousal on Llama-3.1-8B, where a faithful controller would score 1. The model systematically undershoots requested emotional intensity, which puts a number on the qualitative observation of Fazzi et al. (2025). Our experiments trace this to the preference-learning pipeline. Training targets from natural corpora such as EmoBank are neutral-heavy, and the sampled candidates themselves rarely reach extreme affect, so Direct Preference Optimization (DPO) is left with no extreme exemplar to prefer. If instead we cover the target space uniformly and sample a hotter, larger candidate pool, valence gain rises from 0.26 to 0.40 +/- 0.02 (3 seeds) and extrapolation error drops, at only a modest in-distribution cost (EmoBank-test VA distance 0.092 to 0.107). The same recipe reproduces on Qwen3-8B (gain_v 0.44, with in-distribution accuracy preserved). Arousal is harder and less reliable: its gain barely moves on average and swings across seeds (0.14 +/- 0.07, against valence's tight +/- 0.02), because raising arousal needs candidates the base model is reluctant to generate. The evidence indicates that faithful intensity is bottlenecked by the extremity of the candidate pool rather than by the conditioning format.
Read the full abstract

High Energy Physics (Phenomenology)

Not so good $\nu$s for Higgsino dark matter as LZ excess: stringent limits from Super-Kamiokande and IceCube

The LUX-ZEPLIN (LZ) collaboration has recently reported a single nuclear recoil event at a high recoil energy of about 250 keV. This has been interpreted as inelastic scattering of dark matter that is a supersymmetric Higgsino with a mass splitting between the neutral states of a few 100 keV. Such dark matter may be captured at high recoil in the Sun through scattering on heavy elements in it, and annihilate to $W^+W^-$ and $ZZ$, in turn giving rise to a neutrino flux detectable on Earth. Using measurements of atmospheric electron- and muon-neutrino fluxes by Super-Kamiokande and IceCube, we constrain thermal and non-thermal Higgsino dark matter, excluding inter-state mass splittings $\lesssim 557$ keV. This disfavors Higgsino-like interpretations of the LZ event for standard halo velocities.
Read the full abstract

All authors Debajit Bose, Akash Kumar Saha, Deep Jyoti Das, Rinchen Sherpa, Abhijeet Singh, Durba Ghosh, Jaya Doliya, Subhadip Bouri, Biprajit Mondal, Ranjini Mondol, Nirmal Raj, Ranjan Laha

Optimization and Control

Boundary null controllability of weakly coupled structurally damped beams

We prove boundary null controllability in any positive time for a finite system of Euler–Bernoulli beams with possibly non-diagonalizable coupling. For $0<\rho<2$, a Fourier–Jordan reduction yields a vector moment problem with polynomial–exponential modes. Under the Fattorini–Hautus condition, global spectral non-collision, and nonvanishing modal transfer factors, a suitably estimated block-biorthogonal family produces an $H^2$ boundary control. Geometric multiplicities of the coupling matrix determine the minimal control dimension, and Jordan-block lengths determine the degrees of the generalized moments. At $\rho=2$, an exact heat-system reduction gives the same result for $A=A^*\geq0$, despite the time-differential linkage of the reduced boundary inputs.
Read the full abstract

Networking

Real-Time dApps for AI-RAN: Measured Interface Requirements for Inline PHY and Slot-Level Control

Distributed applications (dApps) bring AI to the microsecond-to-millisecond band beside the 5G distributed unit (DU), but every public dApp framework realizes them the same way: an external process that receives an indication and returns a control message. That boundary is right for sensing and advisory workloads. It cannot express a neural receiver that must finish inside a slot, and it cannot hold a scheduling decision the MAC is waiting on. This paper asks what a dApp interface must deliver, in latency and bandwidth, for the AI-RAN use cases now filed under that name to be realized as dApps at all. It treats the coupling between a dApp and the DU as a design axis with two legitimate ends, a C ABI inside the DU process and a protocol-standard E3AP association over SCTP, and places the three OCUDU dApp classes on it. An audited corpus of 39 runtime AI-RAN use cases, sized by 5G NR timing, shows that more than half cannot cross the observer boundary: inline PHY work because an indication has no return path into the same slot, and bounded control because of the tail under load. Measuring the mechanisms each framework actually uses, on a quiet host and with a live cell on the air, shows that every carrier meets a 100 us control deadline on an idle host and that only the in-process paths still do once the DU is running. The asynchronous use cases remain feasible at every position, so the protocol-standard end is kept as a first-class option; what in-process placement adds for them is accountable staleness and an observation-to-decision path with no message on it. These measurements derive the released ABI, a stream for inline work, a validated call for bounded control, and a choice of lease, supervised ring, or portable E3 client for observation, and four dApps of all three classes are validated together on one over-the-air cell.
Read the full abstract

Algebraic Geometry

$F$-nilpotence and Hodge filtrations beyond isolated singularities

We give Hodge-theoretic characterizations of characteristic-zero varieties of open $F$-nilpotent and open weakly $F$-nilpotent type, extending a result of Srinivas and Takagi beyond isolated singularities. We prove one direction unconditionally and the converse assuming the weak ordinarity conjecture.
Read the full abstract

Artificial Intelligence

Understanding the Impact of Model Pruning on Long-Tail Forgetting and Explanation Reliability in Medical Imaging

Model pruning is widely used to compress deep neural networks, reducing memory and computational requirements with minimal impact on aggregate performance. However, its effect on model behavior remains poorly understood, particularly for long-tailed medical datasets where rare but clinically important conditions are underrepresented. Furthermore, it remains unclear whether pruned models preserve reliable explanations of their predictions. To address this gap, we present a systematic study of long-tail forgetting and explanation reliability under model pruning. Across two long-tailed medical imaging datasets, two CNN architectures, four pruning methods, and sparsity levels up to 95%, we evaluate predictive performance, explanation stability, and explanation faithfulness. Our results show that predictive performance exhibits a strong frequency-dependent trend, with lower-frequency classes generally experiencing earlier and larger degradation than higher-frequency classes. In contrast, explanation stability and faithfulness are influenced primarily by the pruning strategy, with gradient-informed methods preserving explanation reliability more effectively under aggressive compression. Qualitative and mechanistic analyses further indicate that explanation degradation is primarily associated with the collapse of class-discriminative gradients rather than the disappearance of feature activations. These findings suggest that model compression should be evaluated beyond aggregate performance. Incorporating class-aware and explanation-aware evaluation reveals failure modes that would otherwise remain hidden, while moderate sparsity levels provide a practical balance between compression, predictive performance, and explanation reliability.
Read the full abstract

All authors Nazish Khalid, Tausifa Jan Saleem, Amal Saqib, Donald C. Wunsch II, Mohammad Yaqub

Probability

Non-asymptotic bounds for the average singular value of a complex Gaussian matrix

Let $G_{d}$ be a $d \times d$ matrix with independent standard complex Gaussian entries, let $\alpha_{\mathbb{C}}(d)$ be the expected average singular value of $G_{d}/\sqrt{d}$, and set $\Delta_d := \alpha_{\mathbb{C}}(d)-\alpha_{\mathbb{C}}(d+1)$. The statistic $\alpha_{\mathbb{C}}(d)$ admits a variational representation as an expected normalized maximum over the unitary group and governs approximation guarantees for the little Grothendieck problem over the unitary group and related unitary registration problems. We obtain a strictly positive lower bound and an upper bound for $\Delta_d$, both valid in every dimension, together with corresponding bounds for $\alpha_{\mathbb{C}}(d)$ around the Marchenko–Pastur limit. These bounds match the sharp leading behavior of complete asymptotic expansions for both quantities, whose coefficients are explicitly computable. The proof combines a three-term recurrence for $Y_d = d^{3/2}\alpha_{\mathbb{C}}(d)$, obtained from its continuous dual Hahn representation, along with singularity analysis of the underlying Laguerre moment generating function.
Read the full abstract

High Energy Physics (Theory)

Logarithmic correction to the entropy of near-extremal higher-curvature black holes

The spherically symmetric sector of broad classes of $D$-dimensional gravitational theories can be effectively described by general two-dimensional Horndeski theories. Exploiting this correspondence, we study the low-temperature dynamics of near-extremal black hole solutions of the parent theories by deriving a universal near-horizon effective description in terms of Jackiw–Teitelboim (JT) gravity. We explicitly show that the logarithmic quantum correction to the entropy, $S_{\rm JT}=\frac{3}{2}\log \left(T/T_{\rm breakdown} \right)$, previously derived for near-extremal black holes in Einstein gravity coupled to matter from the one-loop exact JT partition function, is universal across all models admitting an effective two-dimensional Horndeski description. We determine the general form of the scale at which this correction becomes dominant, $T_{\rm breakdown}$, in terms of the data of the $D$-dimensional theories. We illustrate our general results with charged black holes in Lovelock gravities and regular black holes in Quasitopological gravities.
Read the full abstract

All authors Pablo Bueno, Pablo A. Cano, Robie A. Hennigar, Javier Moreno, Ángel J. Murcia, Guido van der Velde

High Energy Physics (Phenomenology)

Inelastic Singlet-Doublet Fermion Dark Matter in light of the 248 keV LZ event

Recently, the LUX-ZEPLIN (LZ) collaboration reported the observation of a single dark matter (DM)-nucleus scattering event at a nuclear recoil energy of $248\pm23_{\rm stat}\pm23_{\rm sys}$ keV, corresponding to an exposure of 2.84 tonne-year. The absence of events at lower nuclear recoil energies in the predicted spectrum is naturally explained if the underlying process is inelastic DM-nucleus scattering. Motivated by this, we investigate the singlet-doublet fermion DM model, in which the DM consists of two pseudo-Dirac states: the Majorana nature of the lighter state forbids tree level $Z$-mediated elastic scattering identically, while the same states enable inelastic DM-nucleus scattering via $Z$ exchange, with any residual elastic scattering proceeding only through a suppressed Higgs-mediated channel. We further extend the model with a $Z_2$-even scalar triplet, which is responsible both for generating the pseudo-Dirac splitting and for realizing Majorana neutrino masses via the Type-II seesaw mechanism.
Read the full abstract

All authors Debasish Borah, Sujit Kumar Sahoo, Narendra Sahu, Shashwat Sharma

Nuclear Experiment

Precision $\beta$-delayed charged-particle emission spectroscopy at FRIB: Proof of principle with the $\beta$-decay of $^{25}\mathrm{Si}$

We report on the $\beta$-delayed proton and $\gamma$-ray emission from $^{25}\mathrm{Si}$, measured at the Facility for Rare Isotope Beams (FRIB). Low-energy $^{25}\mathrm{Si}$ ions extracted from the Advanced Cryogenic Gas Stopper were implanted into a thin carbon foil surrounded by a compact, highly segmented array of silicon detector telescopes and two high-purity germanium detectors. This setup provides high-resolution charged-particle spectroscopy, establishing a proof of principle for precision stopped-beam decay studies at FRIB. We reconstruct the $^{25}\mathrm{Si}$ decay scheme, resolving new high-energy proton transitions and determining the feeding to excited states in $^{24}\mathrm{Mg}$. The observation of spectral interference patterns enables firm spin and parity assignments for highly excited states in $^{25}\mathrm{Al}$. The $^{25}\mathrm{Si}$ $\beta$-strength distribution is extracted and compared with large-scale shell-model calculations.
Read the full abstract

All authors E. A. M. Jensen, J. M. Eder, P. H. Pedersen, A. Adams, M. J. G. Borge, B. A. Brown, J. Dopfer, H. O. U. Fynbo, B. S. O. Johansson, B. Jonson, M. Madurga, J. S. Nielsen, K. Riisager, C. S. Sumithrarachchi, L. J. Sun, O. Tengblad, L. E. Weghorn, T. Wheeler, C. Wrede

Computation and Language

Climate-ModernBERT: Revisiting Corpus Composition for Domain-Adaptive Continued Pretraining

Natural Language Processing (NLP) in the climate domain requires models to process heterogeneous text sources, including scientific literature, policy disclosures, and synthetic reports. However, how to effectively combine diverse domain corpora during continued pretraining (CPT) remains underexplored. We introduce Climate-ModernBERT, a family of climate-adapted encoder models obtained through continued pretraining of ModernBERT-Base on three climate corpora: academic climate text, climate-filtered web data, and synthetic climate documents. We systematically compare joint continued pretraining on corpus mixtures with parameter-space merging of independently specialized checkpoints. Across nine climate NLP benchmarks, our best model achieves 76.3 average F_1, improving significantly over a vanilla ModernBERT baseline by 2.8 points. Within the climate NLP setting, the results show that academic climate corpora provide the strongest adaptation signal among the evaluated sources, while parameter-space merging improves over joint multi-source training and better preserves complementary information from heterogeneous climate corpora. We release all Climate-ModernBERT variants and training checkpoints to support future research in climate NLP and domain-adaptive pretraining.
Read the full abstract

All authors Yongan Yu, Shantam Raj, Jingwei Ni, Ario Saeid Vaghefi, Dominik Stammbach, Markus Leippold

Computation and Language

Does Syntax Matter? A Graph-Augmented Variational Topic Model for Computational Social Sciences

Topic modeling is widely used in computational social sciences to identify latent themes in large text corpora. Traditional approaches rely on Bag-of-Words representations and generative models such as LDA, while recent methods like BERTopic operate on dense document embeddings. This paper introduces the Structural Contextual Probabilistic Topic Model (SCPTM), an architecture that incorporates syntactic dependency relations into topic inference. SCPTM represents a corpus as a heterogeneous graph of documents and words connected by lexical and syntactic edges, processed through a Graph Attention Network within a Variational Autoencoder to produce probabilistic, mixed-membership topic distributions. We evaluate seven topic modeling techniques (including four SCPTM ablations) across four corpora differing in register and discourse structure. Our framework combines coherence (C_V, C_NPMI), topic diversity, clustering-label alignment (NMI), and phrase-level diagnostics (complementarity and valence gap). Results show that SCPTM's neural architecture yields substantial gains in document-topic alignment over generative baselines, but these gains are attributable to the variational encoder rather than to syntax. Syntax contributes to topic diversity, where graph-augmented variants outperform the no-graph baseline across all corpora, and to descriptor quality: dependency paths capture predicate-argument structures and stance in deliberative registers, while proving redundant in technical and institutional corpora. The valence gap is positive across all variants, but driven primarily by phrase grouping rather than syntactic filtering. We conclude that syntactic encoding matters conditionally: it benefits action-oriented, argumentative texts, but introduces noise in informational or administrative registers.
Read the full abstract

Materials Science

Temperature-Dependent nonlinear optics from first-principles: Second-Harmonic Generation in few-layers MoS$_2$

We present a first-principles real-time approach to study non-linear response of solids at finite temperature. Finite temperature effects are included as renormalization of the quasiparticle energies and a dephasing term proportional to the quasiparticle lifetimes. We evaluate electron-phonon matrix elements from Density-Functional Perturbation Theory for lattice dynamics and then calculate quasiparticles renormalization and lifetime from the Fan and Debye-Waller terms of the electron self-energy. Electron excitations are treated at the independent particle level of approximation. We apply the approach to the second-harmonic generation (SHG) in monolayer and trilayer MoS$_2$. We observe a nontrivial temperature-dependence of the SHG due to a strong crystal-momentum dependent quasiparticle renormalization. From the phonon-mode analysis we find that the coupling with acoustic and shear modes determines the overall crystal-momentum dependence respectively in monolayer and trilayer MoS$_2$. The nontrivial temperature-dependence of the SHG can help rationalize the increase of SHG intensity with increasing temperature observed in monolayer MoS$_2$ at a given laser energy [Adv. Optical Mater. 8, 2000441 (2020)].
Read the full abstract

Human-Computer Interaction

Quantifying the Engagement Trap: Impact of Short-form Video Recommender Systems on Users with ADHD

Short-form video platforms use recommender systems to maximize engagement through highly efficient personalized recommendations. However, the impact of these recommendations on users with ADHD compared to users without ADHD remains underexplored. Through this study, we introduce and operationalize the Engagement Trap, illustrating how recommender systems, while successfully optimizing for engagement, disproportionately disadvantage users with ADHD. This stratified study of 302 participants, recruited via the online platform Prolific, compares experiences between participants with and without ADHD. Our results show that while recommendations are perceived as relevant across groups, participants with ADHD report significantly higher levels of time blindness, post-usage regret, and emotional distress when consuming recommendations. Moreover, we collect feedback for several proof-of- concept, theoretical design interventions for neuro-inclusive design principles. These findings provide quantitative evidence of systemic differences in engagement-optimized recommender systems and highlight the unbalanced negative effects and interactions these systems create for participants with ADHD. We argue for neurodiversity-aware, human-centered design approaches that mitigate such algorithmic harms and support more equitable experiences.
Read the full abstract

Statistics Applications

Two-resolution state modelling of intermittent online gambling activity during Sweden's temporary deposit-cap period

Intermittent online gambling records create two distinct representation problems: calendar-time analyses must retain represented days with no recorded play, whereas analyses of active behaviour must preserve variation within active episodes. We pair calendar-time and active-bout models in a two-resolution analysis of 14.2 million player-days from a single operator during 2019-2023. The calendar-time layer retains assignable represented days in five observed categories and propagates a pre-period reference. The active-episode layer fits a Bayesian mixed-emission hidden semi-Markov model to turnover, playing time, approved deposits and session counts. In the reference four-regime fit, two regimes share a rounded model-implied turnover median of 153.5 EUR, while their model-implied playing-time and approved-deposit medians differ by factors of 4.6 and 5.7, respectively, and their expected total session counts by a factor of 2.8. This separation remains under an ordinary hidden Markov model, two routing perturbations and a residual-dependence sensitivity; the tested turnover-only fits were diagnostically unstable. Over the temporary Swedish deposit-cap period, deposits and turnover per represented player-day were roughly 55% below the propagated reference and active-category occupancy was 7.8% lower. The largest negative deviations in a prespecified Sweden-assigned cohort occurred among high pre-policy deposit-intensity customers. These are operator-specific descriptive comparisons, not causal policy effects.
Read the full abstract

All authors Sam Andersson, Timo Koski, Helga Westerlind, Keenan Lyon, Per Carlbring, Olof Molander

High Energy Astrophysical Phenomena

Two-Phase Structure of Synchrotron-Cooling-Unstable Relativistic Plasma

Using analytic theory, radiative particle-in-cell (PIC) simulations, and fluid simulations, we show that relativistic, synchrotron-cooling, collisionless, high-$\beta$ pair plasmas filament into a two-phase medium. This process occurs through the interplay of the synchrotron cooling instability (SCI) with the synchrotron firehose instability (SFHI). One phase has high plasma~$\beta$ and is infested with small-scale firehose fluctuations, which scatter particles and pin the pressure anisotropy to the firehose-marginal level. The other phase has much lower~$\beta$, causing the suppression of firehose modes and thus allowing large pressure anisotropies. We propose a fluid model for this two-phase plasma, which we use to study the linear and nonlinear evolution of the SCI and SFHI, and to predict the emergence time of the two-phase structure.
Read the full abstract

All authors Agnieszka Wierzchucka, Pablo J. Bilbao, Robert J. Ewart, Dmitri A. Uzdensky, Alexander A. Schekochihin

Cosmology and Nongalactic Astrophysics

Filament-Arm Node Systems (FANS): A new probe of the winding and chirality of the cosmic web

We define Filament-Arm Node Systems (FANS), a new class of cosmic-web objects consisting of multiple filament arms connected to a common node, and use their geometry to construct an object-based test of cosmic chirality. With the cosmic-web node providing a natural reference point, FANS span characteristic scales of ~60 Mpc/h, allowing the coherent winding of their arms around the node to be measured directly. We define cosmic spirality as the amplitude of this winding and handedness as its signed clockwise (CW) or counter-clockwise (CCW) sense. We further introduce an axis-conditioned extrinsic estimator that searches for the projected signature of a coherent three-dimensional axial pattern of the handedness field, corresponding to a preferred axis of winding. Applying these measurements to two distinct SDSS filament reconstructions, and assessing their significance with sign-flip nulls and 27 controlled variations of the FANS construction parameters, we find no significant preference for CW or CCW winding, either globally or around a preferred axis. The data are therefore consistent with parity symmetry, as expected in standard LCDM cosmology. Comparing the lowest- and highest-redshift bins over 0.1<z<0.6, we find that cosmic spirality decreases with cosmic time, a trend that also persists within individual arm-multiplicity classes and may reflect the gravitational straightening of filament arms. These first constraints establish FANS as a new probe of cosmic-web symmetry and large-scale morphology and provide a baseline for future parity tests with the much larger filament samples expected from upcoming surveys.
Read the full abstract