Peer Revenue Publish or profit

Papers

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

All of arXiv

Friday 4 September

Algebraic Geometry

A flatness criterion for pseudo-effective sheaves on compact Kähler spaces

In this paper, we prove that if $E$ is a pseudo-effective sheaf with vanishing first Chern class on a klt compact Kähler space $X$, then, after passing to a finite quasi-étale cover, the reflexive pullback of $E$ is locally free and flat. This extends the flatness criterion of Höring–Peternell, originally established for projective varieties, to the Kähler setting. The proof relies on two main ingredients, both of which are new even in the projective case. The first is a flatness theorem for stable sheaves: we show that a slope-stable pseudo-effective sheaf with vanishing first Chern class is Hermitian flat. This is obtained by combining Hermitian–Einstein theory with the subharmonicity properties of direct image sheaves. The second is a singular Kähler analogue of Simpson's flatness theorem for extensions of locally free Hermitian flat sheaves.
Read the full abstract

Quantum Physics

Fanout Complexity of Symmetric Boolean Functions in $\mathsf{QAC}^0$

Whether $\mathsf{QAC}^0$ can compute $\mathtt{PARITY}_n$ remains open. Computing $\mathtt{PARITY}_n$ is equivalent to implementing $\mathtt{FANOUT}_n$ under $\mathsf{QAC}^0$ reductions. This raises a more general question: for an arbitrary symmetric Boolean function $f:\{0,1\}^n\to\{0,1\}$, what fanout size is necessary and sufficient for computing $f$ in $\mathsf{QAC}^0$? We show that the answer is exactly the transition radius $\rho(f)$: computing $f$ and implementing $\mathtt{FANOUT}_{\rho(f)}$ are equivalent under $\mathsf{QAC}^0$ reductions. In particular, if $\rho(f)\ge n^\delta$ for some constant $\delta>0$, then computing $f$ is $\mathsf{QAC}^0_{\mathrm{f}}$-complete. Combined with Paturi's theorem, our characterization implies that if $\mathtt{PARITY}_n \notin \mathsf{QAC}^0$, then any Boolean function in $\mathsf{QAC}^0$ of approximate degree $n^{1/2+\Omega(1)}$ must be nonsymmetric.
Read the full abstract

Computation and Language

Compression Beyond the Uncompressed: A Two-Stage Training Recipe for Soft Context Compression in RAG

Retrieval-Augmented Generation (RAG) improves knowledge-intensive generation by conditioning language models on retrieved documents, but processing these documents becomes increasingly expensive as retrieval depth grows. Soft context compression reduces this cost by encoding documents into compact continuous representations that can be precomputed and reused across queries. However, many existing methods train compressed models by distilling from a full-context teacher. When the teacher is wrong, such distillation can reinforce its errors, while teacher imitation provides no direct signal for improving beyond the teacher. We propose DEX-Comp, a two-stage training recipe that separates reliable imitation from targeted exploration. Pure Distillation learns only from teacher-correct questions to mitigate error propagation, while Hard Exploration applies outcome-based reinforcement learning to teacher-failed questions to directly optimize answer correctness. Across five open-domain QA benchmarks and retrieval depths from top-$5$ to top-$30$, DEX-Comp at $16\times$ compression outperforms all evaluated compression baselines and surpasses the untuned full-context RAG model in average accuracy, while reducing time-to-first-token by $4.4\times$–$23.7\times$. Evaluations across additional datasets and backbones further demonstrate its generalization.
Read the full abstract

Neural and Evolutionary Computing

Large Language Models with At Most One Spike per Neuron

Leveraging their inherent sparse event-driven computation, spiking neural networks (SNNs) offer a promising path toward energy-efficient large language models (LLMs). Time-to-first-spike (TTFS) coding generates at most one spike per neuron within a time window, yielding extremely low firing rates. However, conventional TTFS SNNs are restricted to specific structures, making it challenging to encode certain blocks in LLM – such as layer normalization and matrix multiplication –using TTFS. To overcome this limitation, we introduce a reference-based strategy specifically to encode the four core LLM components: embedding layers, layer normalization, attention-related operations and dropout. We construct a fully TTFS-based SNN architecture and train it end-to-end. Experiments on modern LLMs like BERT and GPT-2 demonstrate that our approach achieves performance comparable to ANN counterparts on natural language understanding and common-sense reasoning, while a clear gap remains on language modeling perplexity. To the best of our knowledge, this is the first work to scale a spiking LLM to 1.5 billion parameters using TTFS coding. We also report an estimate of spike-related energy; this is a spike-count proxy under an established cost model rather than a measurement on neuromorphic hardware.
Read the full abstract

All authors Zhuoya Zhao, Parsa Omidi, Aref Jafari, Richard Naud

Machine Learning

Beyond Stationarity in Time Series: Discovering Causal Structures and Latent Regimes via Markov Blankets

This paper introduces Regime-aware Constraint-Based and Noise-Based causal discovery with Markov Blankets (RCBNB-MB), a novel causal discovery algorithm for time series that relaxes the common assumption of a single, time-consistent causal structure. Time series are typically observed at discrete time points and often exhibit regime changes that challenge the assumption of a static causal structure, a limitation in many real-world dynamic systems. To address this challenge, RCBNB-MB identifies latent causal regimes, defined as subsets of time points within which a stable causal structure holds. The algorithm follows an iterative strategy that segments the time series into regimes and discovers the causal graph within each regime. By leveraging the Markov blanket rather than direct parents, RCBNB-MB gains robustness to errors in causal discovery and preserves predictive information. We provide theoretical guarantees for RCBNB-MB's ability to recover both regime transitions and causal graphs under reasonable assumptions. Furthermore, we validate its effectiveness through extensive experiments on simulated datasets with known ground truth and real-world IT monitoring data, where taking into account regime shifts is critical. Empirical results show that RCBNB-MB systematically outperforms baseline approaches in accurately detecting regime changes and their associated causal graphs, positioning it as a robust and versatile framework for non-stationary time series analysis.
Read the full abstract

All authors Lei Zan, Charles K. Assaad, Emilie Devijver, Eric Gaussier

Computation and Language

From Vision to Language: Investigating Causal Information Flow in Multimodal Decision-Making

Vision-Language Models are commonly evaluated through their final predictions, but understanding whether these decisions are grounded in visual evidence requires tracing how visual information contributes to language-based decisions. With this purpose in mind, we investigate cross-modal information flow in a video-based generative multiple-choice-like setting by applying a layer-wise causal intervention on video-text attention pathways. We target spatial, causal, and temporal visual reasoning. Our results show that visual information is mainly integrated while the model processes the candidate answer options, which serve as the primary textual grounding sites for the final decision. We further show that nouns play an important role as semantic anchors during multimodal enrichment, while verbs are more relevant when temporal relations are processed. Finally, we identify a distinct pattern in temporal reasoning, suggesting that VLMs struggle to reconstruct sequential information across video frames, but we remark that such fragility may also reflect linguistic biases associated with specific temporal expressions used for defining the relation between events within a scene.
Read the full abstract

All authors Davide Testa, Hugh Mee Wong, Alessandro Lenci, Bernardo Magnini, Albert Gatt

Rings and Algebras

Explicit Super-Linear Algebra over K[{\theta}1, {\theta}2]: A General Berezinian Correction Formula and Multiplicativity Theorem for Arbitrary Block Size

We study super-matrices $M_{m|n}(R)$ over the rank-two exterior algebra $R=K[\theta_1,\theta_2]/(\theta_1^2,\theta_2^2,\theta_1\theta_2+\theta_2\theta_1)$, the smallest supercommutative ring on which the Berezinian's odd-by-odd correction term $BD^{-1}C$ need not vanish. We recall that the supertrace satisfies graded cyclicity, $\mathrm{str}(XY)=(-1)^{|X||Y|}\mathrm{str}(YX)$, so $\mathrm{str}([A,B])=0$ for even $A,B$, and that an explicit $1|1$ matrix over $R$ has $BD^{-1}C\neq0$ surviving to second order in the odd generators. We then prove the extension anticipated in the $1|1$ note: a closed-form Berezinian formula for arbitrary even $X\in M_{m|n}(R)$, $m,n\geq1$ (Theorem 1), given through a single $m\times m$ matrix pairing $\Delta$ built from the odd blocks of $X$; and a fully general multiplicativity theorem $\mathrm{Ber}(XY)=\mathrm{Ber}(X)\mathrm{Ber}(Y)$ for even invertible $X,Y\in M_{m|n}(R)$ (Theorem 2), proved by an explicit trace-cyclicity cancellation. Both results specialize exactly to the known $1|1$ formulas when $m=n=1$, and we verify them on a worked $m=2,n=1$ example. We extend the framework from two to an arbitrary number $r\geq2$ of odd generators, replacing the scalar pairing $\Delta$ by a $\Lambda^2(V)$-valued pairing on the $r$-dimensional space $V$ of odd generators, and identify this pairing, in the scalar case, with a Plücker coordinate of the odd off-diagonal data. We close with a discussion of related work of Khudaverdian–Voronov, and a remark identifying $\Delta$ as a matrix-valued contraction against the natural alternating pairing on the space of odd generators.
Read the full abstract

Materials Science

Interfacial Accommodation as a Candidate Ductility Pathway in Intermetallic-Rich Alloys

Heterophase interfaces are increasingly recognized as active participants in plastic deformation, yet quantitative methods for comparing their strain-accommodation capacity remain limited. Here, we present an atomistic framework for quantifying interface-mediated strain accommodation using molecular dynamics simulations. Interface broadening, roughening, and migration are extracted from atomistic trajectories and combined into an Interface Accommodation Index, with a normalized counterpart accounting for differences in initial interface structure. A weighting sensitivity analysis demonstrates that the relative ranking of interfaces is robust to the specific form of the index. The framework is demonstrated for three experimentally motivated interfaces, Al/Al3Ti, Al/Al9M2, and Al9M2/Al3Ti (M = Fe, Co, Ni), under tensile, compressive, and shear loading. Bulk simulations show that Shockley partial dislocations dominate plastic deformation, with alloy chemistry governing the transition toward mixed-character dislocation networks. Tensile loading produces the greatest interface accommodation, while shear produces comparatively limited structural evolution. The Al9M2/Al3Ti interface exhibits both the highest yield resistance and the largest accommodation response, suggesting that intermetallic-intermetallic interfaces can simultaneously sustain load and redistribute strain. These simulations quantify structural accommodation rather than ductility or fracture directly; their connection to macroscopic ductility therefore requires experimental validation. The framework provides a transferable approach for comparing interface accommodation in multiphase alloys and identifies interface chemistry and crystallography as important design variables for damage-tolerant structural materials.
Read the full abstract

Artificial Intelligence

A Hybrid Predictive Ensemble of Machine Learning and Deep Neural Networks for Early Cardiovascular Disease Risk Assessment

This study introduces an intelligent framework that integrates machine learning and deep neural network ensemble techniques for early detection and prognosis of cardiovascular diseases. The system utilizes real-time physiological data collected from Internet of Medical Things (IoMT) devices, including ECG sensors, heart rate monitors, and blood pressure trackers. To ensure the accuracy and reliability of input data, preprocessing steps such as noise reduction, normalization, and missing value imputation are employed. The most significant health indicators are identified through effective feature selection methods and then processed using optimized classifiers such as Support Vector Machines (SVM), Random Forests, and eXtreme Gradient Boosting (XGBoost), which are combined in an ensemble architecture to improve diagnostic precision. The framework demonstrates remarkable performance in predicting cardiovascular disease risk, achieving higher accuracy, reduced false positives, and enhanced consistency compared to conventional methods. It is designed on a cloud-based infrastructure that ensures scalability and real-time processing for continuous patient monitoring. Experimental evaluation on real-world cardiovascular datasets confirms the framework's efficiency in early-stage risk assessment and clinical decision support. The results highlight the potential of combining traditional machine learning and deep learning paradigms to achieve proactive healthcare management and improve patient outcomes.
Read the full abstract

Nuclear Theory

Signals for Nuclear Solid-Liquid Phase Transition in Clustering Nuclei

The nuclear phase diagram remains a fundamental challenge, where uncovering how the cluster degree of freedom evolves with temperature and energy is essential to understand the intriguing crossover between cluster, nucleon, and nuclear astrophysics. A new first-order phase transition, the nuclear solid-liquid phase transition (SLPT), emerges spontaneously from nucleonic degrees of freedom in low-energy heavy-ion collisions. Clear signals of the SLPT are identified as a region of negative heat capacity in smaller systems (e.g., $^{12}\text{C} + ^{12}\text{C}$ and $^{16}\text{O} + ^{12}\text{C}$) and a caloric plateau in larger systems (e.g., $^{28}\text{Si} + ^{12}\text{C}$ and $^{40}\text{Ca} + ^{12}\text{C}$), along with prominent behaviors in the number of fragment species and information entropy. Moreover, the pairwise distance probability density function provides a direct spatial fingerprint of the melting, where long-range crystalline $\alpha$-cluster peaks vanish above the transition temperature, giving way to a liquid-like distribution. This new phase transition unifies the many-body dynamics of finite constituents, extending from nanoscale clusters to nuclear systems. Beyond advancing nuclear structure and phase diagrams, these insights offer profound implications for astrophysics, notably linking structural evolution to astrophysical phase transitions via nuclear cluster dissolution in supernovae.
Read the full abstract

All authors Xi-Guang Cao, Yong-Hao Jin, Chun-Wang Ma, Wen-Bo Liu, Yu-Gang Ma

Number Theory

Counting level-raising congruences using modular representation theory

We introduce a new method for studying mod-$\ell$ congruences between eigenforms through the modular representation theory of $\mathrm{PGL}_2(\mathbb{F}_p)$. When $p\equiv \pm 1 \pmod{\ell}$, we use this theory to construct and describe extra structures on spaces of modular forms with $\Gamma_0(p^2)$-level at $p$ and a fixed mod-$\ell$ Galois representation. The structural results we obtain can be viewed as a refinement of classical level-raising theorems since they not only allow us to prove the existence of congruences, but also to count the number of such congruences. Our methods work equally well in the residually irreducible and residually reducible cases, allowing us to prove several new instances of congruences between Eisenstein series and cuspforms (as well as independently rederiving classical results of Mazur and more recent results of Lang–Wake). Notably, our approach proves these results without computing constant terms of Eisenstein series, without Galois deformation theory and $R=\mathbb{T}$ theorems, and without the Jacquet–Langlands correspondence.
Read the full abstract

Computation and Language

A Human-in-the-Loop Framework for AI-Assisted Scoring in Large-Scale Writing Assessment

The integration of artificial intelligence (AI), particularly large language models (LLMs), into educational assessment has opened new opportunities to enhance the efficiency and scalability of grading processes. This study presents the design and validation of an AI-assisted scoring framework for written responses in a large-scale national assessment. The proposed approach focuses on short written texts of approximately 150-200 words and incorporates a human-in-the-loop strategy to preserve assessment quality while reducing manual workload. The study is grounded in a real operational context, using data from two recent editions of a nationwide test, each comprising approximately 5,000 student responses. We analyze the alignment between AI-generated scores and human raters across multiple rubric dimensions, as well as the impact of the proposed decision flow on pass/fail outcomes. Results show moderate to high agreement between the model and human evaluations in most dimensions, supporting the feasibility of AI assistance in this setting. Moreover, the proposed correction workflow identifies cases where human review is most valuable, enabling a more efficient allocation of expert effort. The findings suggest that AI-assisted scoring can be safely integrated into large-scale assessment processes only when combined with carefully designed human oversight. The paper concludes by discussing practical implications for deployment in national assessment systems and outlining future research directions, including longitudinal monitoring of model-human alignment and the analysis of potential cognitive bias introduced by AI-supported review workflows.
Read the full abstract

All authors María Eugenia Curi, Germán Capdehourat, Isabel Amigo, Magdalena Romano, Rosana Serra, Adrián Silveira, Andrés Peri

Exactly Solvable and Integrable Systems

A note on the integrability status of a 'mysterious' case of a quadratic Hamiltonian

We present a simple algebraic proof of meromorphic non-integrability of a 'mysterious' case of a quadratic Hamiltonian studied by Sokolov and Wolf. The result is obtained by applying the Morales-Ramis method, based on differential Galois theory.

Artificial Intelligence

SciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding

Scientific papers require models to integrate evidence across text, equations, figures, tables, code, and datasets while preserving its provenance. Beyond answer correctness, scientific reading requires verifiable outputs from operations such as evidence localization, definition extraction, and consistency checking. We introduce SciDocBench, a workflow-centered benchmark targeting these operations through 124 expert-authored and difficulty-screened questions across seven research-assistant capability groups, 19 subtasks, and five scientific domains. Each question is instantiated in four matched settings formed by pairing its bilingual variants with the All Images First and Markdown Interleaved document representations, yielding 496 evaluation instances. The strongest evaluated model, Claude-Opus-5, scores 62.6 out of 100, with remaining gaps in evidence localization, structured information extraction, cross-document synthesis, and robustness to document representation. To convert these diagnostics into scalable training signals, we introduce SciDocIR, a structured representation of scientific document objects, layout and cross-reference relations, and provenance. Using SciDocIR, we construct SciDocDataset, which contains 4K supervised fine-tuning instances and 10K reinforcement-learning instances, built on 14 verifiable training subtasks. Post-training Qwen3.6-27B on task-aligned data improves its SciDocBench score from 40.03 to 45.33 with supervised fine-tuning and to 45.74 with subsequent reinforcement learning. Both adapted models preserve DocVQA and InfoVQA performance and improve ChartQA accuracy over the original model by 0.80 and 3.40 points, respectively. Together, SciDocBench, SciDocIR, and SciDocDataset connect capability diagnosis with verifiable training-data construction for scientific-document assistants.
Read the full abstract

All authors Shenxi Wu, Yuhong Liu, Haosong Zhang, Tongjin Zou, Yanxun Zhang, Gaochang Chen, Dun Liang, Jiaqi Wang, Zhecan James Wang, Yuhang Zang, Dahua Lin

Image and Video Processing

Cross-dataset transportability of pediatric chest X-ray deep learning across three countries: discrimination, calibration, operating-point failure, and limited-label recovery

Background and Objective: External evaluation of medical-imaging AI is often collapsed into discrimination. We evaluated a computational protocol that separately tests discrimination, probability calibration, fixed operatingpoint transport, shortcut-associated signal, and limited-label recoverability for pediatric pneumonia classification across datasets from three countries. Methods: After exact-duplicate removal, 5,824 Guangzhou radiographs supported leakage-controlled source development and internal testing. A frozen three-seed DenseNet121 dual-view ensemble was evaluated zero-shot on BDCXR-3257 from Bangladesh (n = 3, 257) and an untouched harmonized VinDr-PCXR/PediCXR test cohort from Vietnam (n = 1, 077). Matched seed-42 variants tested architectural robustness. Secondary BDCXR analyses used a fixed 651-image adaptation pool and 2,606-image hold-out; 163, 326, and 651 labels represented 5%, 10%, and 20% of complete BDCXR. Results: Internal AUROC was 0.976 with 95.1% sensitivity. BDCXR and VinDr-PCXR AUROC were 0.798 and 0.742, while frozen-threshold sensitivity fell to 6.2% and 0%. Source-to-BDCXR AUROC degradation occurred for a full-image baseline (0.961 to 0.749), ungated dual-view model (0.977 to 0.766), and gated MixStyle model (0.966 to 0.789). With 163 BDCXR labels, Platt recalibration preserved AUROC while increasing held-out sensitivity to 88.3%, but specificity was 47.9% and the alert rate was 78.5%. Two hundred repeated 163-label fits confirmed sensitivity recovery but substantial specificity variability. Conclusions: Cross-dataset shifts across countries affected ranking, probability alignment, and source-defined decision behavior differently. Transport studies should evaluate these components separately and quantify the operational burden of apparent recovery.
Read the full abstract

Computation and Language

NS-ST-GraphRAG: Neuro-Symbolic Spatio-Temporal GraphRAG for Literary Knowledge Processing

Long-form literary narratives pose a distinctive information-processing challenge for retrieval-augmented generation: relevant evidence is distributed across chapters, relations evolve over narrative time, and correct answers may depend jointly on temporal, spatial, and relational constraints. We propose NS-ST-GraphRAG, a neuro-symbolic spatio-temporal GraphRAG framework that integrates ontology-guided extraction, deterministic constraint checking, dual temporal coordinates, spatial scene attributes, and dynamic sub-graph retrieval. Instead of retrieving from a single corpus-level graph, the framework selects the graph state valid for the temporal and spatial scope of a query and grounds generated answers in traceable evidence. We further introduce Red-Chamber-QA, to our knowledge the first open multi-hop question-answering benchmark for classical Chinese literature, with time-, space-, and general-question categories, per-part evidence spans, and deterministic shortcut controls. On a 120-question held-out split, NS-ST-GraphRAG achieves mechanical answer reproduction of 0.733 versus 0.675 for the frozen window baseline and 0.083 for a closed-book model (McNemar exact p = 0.092, directionally favorable but not significant); semantic-judge accuracy is 0.866 versus 0.850. The pre-specified constrained-category condition of H2 is not supported by the delivered comparison. These results show how temporal graph representation, constrained extraction, and auditable evaluation integrate into a unified framework for verifiable knowledge processing over long-form narrative.
Read the full abstract

Machine Learning

From 80x to 385x: A Best-Matching-Unit Search at the L2 Roof, Measured Against a Symmetrically Tuned Baseline

Comparisons between GPU implementations are usually asymmetric: one side is tuned by its author, the other is run as found. I report a programme that tuned both a novel SOM algorithm (SparseBin) and the baseline algorithm it was being compared to (cuSPARSE). The best-matching-unit search that dominates self-organizing map training was tuned through four levers - tile size, tile-membership clustering, neuron-axis chunking and vectorised loads - reaching 5.6-10.1x per epoch over the previously published configuration at map sizes from 32x32 to 512x512, and lifting the margin over the CUDA implementation behind our earlier MEDLINE atlases from ~80x to ~385x. cuSPARSE, the implementation SparseBin is compared against, received every lever with an analogue on its side, and became 2-3x faster in the process. The tuned kernel pressed the L2 bandwidth roof at 77% of peak with every other unit at 40-65%, bounding any further lever at ~1.3x - a terminal result rather than a waypoint, and every untested lever was either capped by that bound by construction or measured null.
Read the full abstract

Strongly Correlated Electrons

Environment-assisted transport in a strongly correlated boundary-driven Fermi-Hubbard chain

We study steady-state transport in a one-dimensional Fermi-Hubbard chain coupled to particle reservoirs at the boundaries and to local dephasing baths at each site, using the time-evolving block decimation (TEBD) method to solve the Lindblad master equation. In the absence of dephasing, the current exhibits two well-separated maxima as a function of the boundary driving rate, reflecting the distinct charge and spin energy scales of the strongly correlated regime. Upon introducing dephasing, we find two distinct dephasing-induced transport-enhancement regimes, in contrast to the single enhancement previously reported for spinless fermions. Analysis of the non-equilibrium steady state in the Hamiltonian eigenbasis reveals that the two regimes originate from distinct dephasing-induced redistribution processes: the first involves redistribution within the uppermost Hubbard band, while the second involves transitions between Hubbard bands. Our results demonstrate how many-body correlations shape the interplay between coherent driving, dephasing, and quantum Zeno physics in strongly correlated open systems.
Read the full abstract

Machine Learning

MomentQuant: an even more minimalist interval method with linear time complexity for time series classification

Time series data is very common in many real-world applications and in numerous domains, with increasing interest for automated information extraction using machine learning. One of these subfields is time series classification, which consists in assigning a label to each new, unseen time series. Many algorithms have been developed over the past decades, with the trade-off between predictive performance and computational cost being consistently discussed. Quant, an interval-based algorithm extracting quantiles from recursive, fixed, dyadic intervals, was shown to achieve high accuracy, while being very fast. We propose two changes to make this algorithm even faster. The first one is a better optimized implementation of the exact same algorithm. The second one is to derive approximate quantiles, using the Cornish-Fisher expansion, instead of exact quantiles. This change removes the necessity to sort the time series, leading to a smaller computational complexity. We call this novel algorithm MomentQuant. We provide evidence that our implementation of Quant is faster than the original one, and that MomentQuant is even faster than our implementation of Quant, at the cost of a tiny decrease in predictive performance. These improvements are especially relevant for real-life applications, where inference is performed much more often than training.
Read the full abstract

Chemical Physics

Polarizable Embedding for Photoelectron Spectroscopy and Photoelectron Circular Dichroism in Solution: Core-Level Ionization of Aqueous Alanine

We present a polarizable quantum-mechanics/molecular-mechanics formulation for the calculation of molecular photoionization observables in solution. The approach couples the static-exchange density-functional-theory treatment of bound and continuum electronic states with a fully atomistic polarizable embedding described with the fluctuating-charge force-field. The fluctuating charges are determined self-consistently with the ground-state quantum-mechanical density and are subsequently included in the static-exchange Hamiltonian to account for the outgoing electron. The numerical behavior of the coupled scheme is validated and applied to the X-ray photoelectron spectrum and photoelectron circular dichroism of zwitterionic L-alanine in aqueous solution, using an ensemble of configurations extracted from molecular dynamics. Calculated spectra reproduce experimental profiles remarkably well, demonstrating that atomistic solvent structure and environmental polarization must be treated together to reliably model photoelectron observables in aqueous solution.
Read the full abstract

Earth and Planetary Astrophysics

Stability of circumbinary planets: the role of binary properties and migration scenarios

Among the thousands of exoplanets detected to date, only a very small fraction are classified as circumbinary planets, and this number becomes negligible when considering low-mass planets. This rarity may partly result from observational biases, but also from the challenging dynamical environment of binary systems, which can prevent the long-term stability of planetary orbits. It is therefore essential to investigate the conditions that enable the formation and survival of stable circumbinary planets, particularly in the low-mass regime. To this end, we performed N-body simulations coupled with a planet migration prescription to study the post-formation dynamics of two 10 Earth-mass circumbinary planets in a protoplanetary disc. Based on more than 1,000 simulations, we find that inward migration can, in some cases, be halted through resonance capture with the binary. In contrast to some previous purely N-body studies, these configurations appear to remain stable over long timescales. We further find that the binary mass ratio $q_B$ and eccentricity $e_B$ strongly influence the stability of multi-planet systems. Within the explored parameter space, some regions support long-term stability, whereas others are highly unstable. These stability regions are also sensitive to the migration timescales, and therefore to disc properties and planetary masses. Finally, our simulations suggest that systems in which two planets enter resonance while migrating together are more likely to form stable multi-planet configurations, particularly around highly eccentric binaries. These results provide a first step toward identifying binary star systems that are most promising for hosting multiple low-mass circumbinary planets.
Read the full abstract

All authors Romain Grane, Nicolás Cuello, Mario Sucerquia, Emmanuel Gianuzzi, Milenne Ávila-Bravo, Carolina Charalambous, Amaury H. M. J. Triaud

Robotics

A Schema Bounded Language Model for Refining Robot Policies Without Destabilizing Local Learning

This paper addresses navigation by composite heterogeneous robots in a decentralized system when policy reasoning and local control operate at different update levels. In a NetLogo–Python implementation, three robots share motion dynamics but use different LLM backends. Each robot independently combines a large language model (LLM) policy agent, an Upper Confidence Bound (UCB) bandit, and a Double Deep Q-Network (Double DQN) controller; no central LLM generates team actions. LLM inference is confined to round-level policy generation and refinement rather than tick-level action selection. The robots perform cross-LLM communication through a shared round summary containing policies, outcomes, and learning feedback. UCB performs refinement-mode selection, and the policy-conditioned Double DQN performs tick-level action selection from navigation variables, active policy parameters, and the LLM action prior. Each of the four configurations was evaluated over 30 rounds. In the fixed simulation, the complete configuration reached the goal in all 90 correlated robot–round records and achieved the lowest median completion time (42 ticks) and P90 (73.2 ticks); its median was 25.0–39.1% lower than those of the other configurations. These observations provide descriptive, configuration-level evidence from the evaluated configurations.
Read the full abstract

All authors Chongwen Dong, Mithun Paul Saint-Germain, Pinjari Asif, Carlo R. daCunha

Game Theory

Strategic Facility Location in Euclidean Spaces

The strategic facility location problem is defined as follows: $n$ agents report their location in a metric space, and the objective is to design a mechanism deciding the (possibly randomized) location of a facility such that agents have no incentive to lie about their position. We focus on the egalitarian cost, which means that the goal of the mechanism is to minimize the expected maximal facility-agent distance. Meanwhile, mechanisms must be truthful (or strategyproof): no agent may decrease their expected distance to the facility via lying on their location. Designing truthful mechanisms minimizing the approximation ratio is a well-studied problem, and the optimal solution is known for the real line. We focus in this paper on higher dimension Euclidean spaces, for which gaps remain between the best known lower and upper bounds. We first show that, maybe counter-intuitively, the problem is easier for two agents on the plane rather than on the line: the mechanism can exploit the additional dimension to prevent more efficiently agent lies. Based on this intuition, we devise lower bounds for $\mathbb R^d$ asymptotically matching the best known approximation factor of $2$ for large $d$. We also provide novel mechanism ideas, improving over the best known algorithms on the plane, and when the agents belong to $\mathbb R^d$ but the facility may use an additional dimension.
Read the full abstract

Combinatorics

Real-rootedness and interlacing for parking functions and Chow polynomials

We prove real-rootedness for the Chow polynomials of the noncrossing partition lattices by transferring tieless parking functions to finite-alphabet Smirnov words and applying an interlacing-preserving transition of M.~Leander. We also derive a triangular recurrence for peaks and ties and identify the peakless-tieless descent polynomial as the Narayana polynomial. For the toric $g$-contributions of Ehrenborg–Hetyei–Readdy, we exhibit a fixed-row common interlacer and establish real-rootedness of all nonnegative row sums. Individual real-rootedness follows in particular; Q.~Xiao also proved it by a different differential recurrence. We also give a second proof of the individual statement, by finite Schur–Szegő convolution, that does not use the common interlacer. These results prove Conjecture~4.2 of Xiao and Conjecture~11.2 of Ehrenborg–Hetyei–Readdy, with consequences for weakly 123-avoiding parking functions. We also prove real-rootedness for the image-size polynomial on all parking functions and for the ascent and descent polynomials of four two-pattern-avoiding classes.
Read the full abstract

Materials Science

Semiclassical thermoelectric transport in disordered Dirac electron system Ag2Te

We investigated the thermoelectric effects of the Dirac electron system Ag2Te under magnetic field. Our anal-ysis based on the Boltzmann semiclassical model associated the disorder with the unconventional magnetic field responses such as linear magnetoresistance, linear Nernst effect, step-like Nernst effect, and sign change in Nernst effect. The analysis also revealed the impurity band near the Fermi energy. We simultaneously clari-fied the serious impact of the thermal Hall effect on the measurement of the Nernst effect, and we proposed the definitive solution. Our careful measurement and analysis will be the standard for the thermoelectric study under magnetic field.
Read the full abstract

All authors Kentaro Kuga, Keisuke Hirata, Daiki Goto, Ryogo Ishihara, Masaharu Matsunami, Tsunehiro Takeuchi

Materials Science

Anharmonic Lattice Dynamics and Anisotropic Electron-Phonon Coupling in Quasi-1-Dimensional Charge Density Wave Ta2NiSe7

The microscopic origin of charge density wave formation in quasi one dimensional Ta2NiSe7 remains actively debated, particularly regarding the relative contributions of Fermi surface nesting, electron phonon coupling, and lattice instabilities. Here, we combine temperature and orientation dependent polarized Raman spectroscopy with first principles calculations to uncover the anisotropic electronic and lattice interactions governing the CDW state in Ta2NiSe7. Heat capacity and electrical transport measurements identify an incommensurate CDW transition at 61 K. Raman spectroscopy reveals pronounced in plane anisotropy, with e ph coupling strength along intrachain b axis exceeding five times that along interchain c axis, whereas lattice anharmonicity is enhanced by threefold along c axis. First principles calculations identify Ta2 Se octahedral vibrations and Ta Se electronic states near the Fermi level as dominant channels mediating the anisotropic e ph interaction. Exceptionally strong and directional e ph coupling along b axis establishes lattice driven electronic instability as the primary mechanism underlying CDW modulation and highlights the dominance of intrachain interactions in strongly coupled Ta2NiSe7. Despite this strong coupling, the CDW remains incommensurate, indicating that lattice anharmonicity provides an additional degree of freedom. Enhanced anharmonicity along c axis suggests that anisotropic phonon phonon interactions reshape the free energy landscape and contribute to stabilizing incommensurate phase. These findings reveal a cooperative interplay between anisotropic e ph coupling and lattice anharmonicity in governing CDW formation in low dimensional quantum materials.
Read the full abstract

All authors Prithwija Mandal, S. Nanthini, Aditya Singh, Kewal S. Rana, Dibyendu Dey, Kanishka Biswas, Ajay Soni

Solar and Stellar Astrophysics

Can isolated binaries form unequal-mass binary black-hole mergers with a high-spin primary black hole?

GWTC-5 has revealed a subpopulation of merging binary black holes (BBHs) with a high-spin primary black hole (BH) and possibly unequal BH masses. GW241110 additionally exhibits a large spin-orbit misalignment, which suggests a hierarchical-merger origin. However, other formation scenarios are possible or even likely, especially for events without constraints on spin-orbit misalignment. As an alternative to hierarchical mergers, we investigate whether binary evolution can produce unequal-mass BBH mergers with a high-spin primary BH. Rather than performing comprehensive population-synthesis calculations, we examine the evolutionary pathways of forming merging BBHs and assess their uncertainties. We identify two possible pathways for producing unequal-mass BBHs with a high-spin primary. In initially wide binaries, mass-ratio reversal can make the tidally spun-up second-born BH both more massive and more rapidly rotating than the first-born BH; alternatively, in an initially close, unequal-mass binary, the primary star may evolve chemically homogeneously, while the secondary star evolves normally, producing a high-spin first-born BH that is more massive than its companion. Generally, large spin-orbit misalignment can be produced by large natal kicks or tertiary-induced nodal precession and/or Zeipel-Lidov-Kozai oscillations. We conclude that hierarchical mergers are not uniquely required to produce unequal-mass BBHs with a high-spin primary BH, although each isolated-binary pathway faces important theoretical constraints. Future detections of more merger events with primary BH spins around 0.7 would discriminate between binary evolution and hierarchical merger origin.
Read the full abstract

Graphics

GradRig: Differentiable Weights for Skinned Gaussian Splat Deformation

Skinned deformation is a common framework to turn a 3D shape from its rest pose into a dynamic pose through the deformation of a coarser kinematic structure, called rig. When applied to a 3D mesh, this rig only needs to displace vertices to deform the polygons that connect them. However, when deforming 3D Gaussian Splats, which do not provide connectivity information, rigidly transforming points is not enough to prevent the creation of holes when stretching shapes. In this paper, we use the spatial gradient of skinning weights to provide a full mesh-free deformation pipeline for Gaussian Splats, that more accurately stretches splats while remaining fully compatible with real-time rendering capabilities, which we demonstrate in a WebGL viewer. We present how we evaluate these gradients when the user creates the rig structure and propose an optional adaptive resampling scheme to split up splats that still produce artifacts.
Read the full abstract

Machine Learning

Coarse-Graining Hidden Representations: Unsupervised Neuron Selection via Mapping Entropy

Overparameterized neural networks carry far more hidden units than a task nominally requires, raising the question of which neurons are essential and whether that distinction is legible in the representation itself, without labels or gradients. We cast neuron selection as the problem of coarse-graining the hidden layer by retaining a subset of its neurons, and score each putative selection by the mapping entropy (ME). This quantity measures the loss of discriminatory power inherent in discarding part of the network neurons, and the selection that minimises the ME is taken as particularly informative. This criterion is fully unsupervised, in that it depends only on hidden-activation statistics. In teacher-student networks, ME optimisation recovers the minimal teacher-consistent representation and retains extra units in proportion to the hidden layer's residual variability; in a non-linear Gaussian process task, it selects coherent functional-class mappings whose preferred class shifts across training. On this task and on translation-augmented MNIST, ME-selected subnetworks outperform random subsets of equal size, most clearly under strong compression - linking configurational distinguishability to predictive performance.
Read the full abstract

All authors Margherita Mele, Andrea Castagna, Roberto Menichetti, Raffaello Potestio, Alessandro Ingrosso

Machine Learning

Evaluating and Recalibrating Probabilistic Classifiers from One-Bit Threshold Queries

Standard calibration evaluation and recalibration methods assume access to predicted probabilities. We study one-bit threshold access, a setting in which probabilities are hidden and the evaluator observes only whether the predicted probability exceeds a single chosen threshold per sample. Under this restriction, we show that classifiers can nevertheless be evaluated and recalibrated without reconstructing their hidden scores. We derive a consistent estimator of True Calibration Error, unbiased estimators of strictly proper scoring rules, and a one-bit objective for Temperature Scaling whose expectation matches the standard log-loss objective. Across these settings, we quantify the statistical cost incurred by replacing standard probability-access procedures by our one-bit estimators. Finally, we instantiate the framework for black-box LLMs, where the logit_bias API parameter allows us to realize the required threshold queries in zero-shot binary question answering.
Read the full abstract

All authors Roman Plaud, Antoine Saillenfest, Matthieu Labeau, Thomas Bonald, Willem Waegeman