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

Tuesday 15 September

Optimization and Control

Integrating Wildfire Risk into Robust Forest Management Optimization

Increasingly severe and unpredictable extreme weather events-such as high-intensity forest firesthreaten long-term ecosystem service provision, driving a critical reassessment of current forest management planning. While planning frameworks exist, few incorporate optimization methods that account for the uncertainty and variability introduced by such extremes. Meanwhile, process-based models have advanced our understanding of fire impacts on forest dynamics and ecosystem services, yet their potential to inform strategic management decisions remains largely untapped. Focusing on intensive plantation forests, we propose a combined approach that uses fire simulation within a growth model to optimize forested landscape management under increasing fire risk. We demonstrate that incorporating fire behavior and its uncertainties significantly alters management outcomes, favoring diversified portfolios that balance high-yield but fire-prone regimes with lowerrisk alternatives. While diversification improves robustness for objectives like timber provision, it does not guarantee complementary benefits for carbon storage. Our method exposes the limitations of current management regimes and identifies key areas for improved planning to better mitigate fire risk and strengthen long-term resilience.
Read the full abstract

All authors Clémence Labarre, Denis Loustau, Jean-Christophe Domec, Mengistie Kindu, Thomas Knoke

Robotics

TEMPO: Learning Temporal Context for Dynamic Robot Manipulation

Vision-language-action (VLA) models have achieved impressive performance in quasi-static manipulation, but struggle in dynamic manipulation tasks because they operate on a single observation at inference time. We identify two representational failures that underlie this limitation. The first is motion ambiguity, where a single observation does not include scene dynamics and therefore cannot anticipate the future state of moving objects. The second is state aliasing, where visually similar observations from different points in a task require different actions. We argue that these failures persist regardless of model scale and inference latency, showing that the bottleneck is missing temporal context rather than model capacity. Based on this insight, we propose TEMPO, which augments a pretrained VLA with two temporal inputs: a motion summary extracted from a frozen video foundation model to resolve motion ambiguity and a compact proprioceptive history to resolve state aliasing. TEMPO requires no modification to the backbone and adds minimal compute overhead at training or deployment. Across four dynamic manipulation tasks, it improves Bottle Handover success from 44% to 74% and is the only method that solves state aliasing. Probing and ablation studies confirm that each temporal signal independently addresses its corresponding failure. We further release TEMPO-Bench, a benchmark of over 50k annotated frames for evaluating motion-aware robot perception in both regression and multiple-choice formats. Project Website: https://tempo-robot.github.io/
Read the full abstract

All authors Zhenyang Feng, Jimin Heo, Erik B. Sudderth, Unnat Jain

Numerical Analysis

Mixed precision solvers for the all-at-once Runge–Kutta discretization of the heat equation

We study the effect of mixed precision on the numerical integration of the heat equation discretized with a Runge–Kutta method in time. A full space-time discretization is applied, which results in a very large and sparse linear system to be solved for the numerical approximations and the Runge–Kutta stages of all time steps. The linear system is solved by applying a suitable preconditioned iterative method that can be run in parallel. In order to speed up the solution process, the preconditioner is applied using a mixed precision framework. Sequential results show the robustness of the preconditioner, even when applied in mixed precision. Finally, we present numerical evidence of the improved performance of the mixed precision strategy when applied in a parallel environment, achieving up to a 50% reduction in CPU time.
Read the full abstract

All authors Santolo Leveque, Luca Bergamaschi, Ángeles Martínez, Erin Carson

Solar and Stellar Astrophysics

Triple coronal hard X-ray source observed by STIX during a failed filament eruption

{Observations reveal hard X-ray sources in the solar corona with morphologies and locations distinct from typical solar flares. One expected mechanisms for hard X-ray (HXR) production in the solar corona is eruptions of various types. Failed eruptions are promising targets for detecting non-flare-related HXR sources because most of their energy is expected to be dissipated in the corona. } {We investigated the SOL2022-0 to determine the nature of coronal HXR sources associated with the event.} {We used Solar Orbiter's STIX, EUI, and SDO AIA observations. We reconstructed STIX images using the MARLIN algorithm. We used EUI data to cross-check the view of the event from two vantage points. The AIA images enabled kinematic analysis and reconstruction of differential emission measure (DEM) maps. We then used the AIA DEM to predict X-ray emission maps and compared them with STIX images. } {We find temporal and spatial correlations between the failed eruption and HXR sources. During the braking of the filament eruption, we observed non-thermal HXR emission from coronal sources along the path of the eruption. Thermal HXR emission was concentrated in three coronal sources formed with two separate mechanisms: chromospheric evaporation and direct heating. Directly heated sources are the result of the interaction of the eruption with the overlying magnetic field, and were observed to cool for an extremely long time.} {The HXR emission sources related to the interaction between the eruption and the overlying magnetic field might be typical structures in all flares. However, they appear to be weak, since their emission is only 5-20~% of that from the flare coronal source. If footpoint sources were not occulted, then these weak coronal emission sources would not be visible due to the low dynamic range of HXR telescopes.}
Read the full abstract

All authors Tomasz Mrozek, Marek Stęślicki, Sylwester Kołomański, Krzysztof Barczynski

Performance

Strong aggregation of the Markov chains associated with matching models based on the automorphism group of their compatibility graphs

We extend the analysis of strong aggregation to general compatibility graphs, focusing on item counts rather than positions, and exploring generalized greedy matching disciplines. We prove that under a condition of automorphism-based transition consistency, the associated Markov chain is strongly aggregable for an arbitrary graph with a non-trivial automorphism group. Furthermore, we extend our analysis to non-greedy matching disciplines, distinguishing scenarios where compatible items can or cannot coexist within the same state. This result is illustrated with a simple compatibility graph with a rich automorphism structure: the odd rings. For all scenarios, we investigate the strong aggregation properties of the resulting Markov chains. This work enhances the theoretical understanding of lumpability in stochastic matching models and provides a foundation for analyzing complex graph structures.
Read the full abstract

Computation and Language

Reduplicative constructions in Mandarin: Socio-emotional profiling through distributional semantics

Mandarin Chinese has two productive reduplicative constructions that repeat either two-character base words or their constituents (e.g., `in good health', `discuss a bit'). Their varied meanings have been described as realizing plurality, valence coloring, sound symbolism and pragmatic functions. The aim of this study is twofold. A first goal is to clarify whether it is possible to come to a more precise understanding of the variegated semantics of Mandarin reduplication by using word embeddings from distributional semantics. A second goal is to explore how useful embeddings are for understanding the details of a semantically complex word-formation process. We show that the embedding space recovers the semantic and grammatical properties of reduplications previously identified in the literature, validating Tencent embeddings for morphological investigation. Semantic profiling revealed that reduplicative constructions are often strongly represented on multiple dimensions. The two patterns exhibit clear semantic and pragmatic differentiation in distributional space. Procrustes analysis clarified that the overall organization of the base-word space is largely preserved in the reduplication space, with local mismatches highlighting regions of discourse-pragmatic reorganization. Taken together, these results show that high-dimensional word embeddings can recover established linguistic generalizations, and capture the semantic versatility of Mandarin reduplication and constructional transparency.
Read the full abstract

Computer Vision

Measuring Annotation Efficiency for Handwritten Devanagari Recognition: Sample-Complexity Curves for Four Pretraining Regimes

To train handwritten text recognition systems we need word images and their corresponding transcriptions, and these transcriptions are produced manually. For a script that can be read by only a small number of specialists, this manual transcription is a limitation, because the trained models are supposed to save the time of those same specialists. A relevant question therefore arises: how many transcriptions are needed before a recogniser becomes useful, and how much of that cost can pretraining remove? In this study the answer is measured directly for handwritten Devanagari. We keep the recogniser, optimiser and evaluation protocol the same and change only the number of real transcribed words used for fine-tuning across nine budgets from 10 to 4,000 and four initialisation regimes, with six seeds at every point. The resulting curves are then converted into annotation-equivalent terms. A CER of 0.50 is reached by supervised synthetic pretraining using only 81 transcribed words, whereas random initialisation requires 355, which gives a label multiplier of 4.40 [3.56, 4.99]. There is a zero-shot reference point as well: with no real transcribed words at all, this pretraining is worth about 136 of them. This advantage gets smaller as the target accuracy improves, and at the most demanding target we measure, it cannot be distinguished from no saving at all. A fourth arm in which only the encoder is transferred separates the effect of the pretraining method from that of transfer scope, and masked image modelling is observed to transfer negatively over a bounded range of budgets. We emphasise that the scarcity in this study is constructed by subsampling a large corpus.
Read the full abstract

Computer Vision

TecoPrompt: Temporal-Conservative Prompt Learning for Vision-Language Models

Prompt learning adapts vision-language models, such as CLIP, by adjusting a small set of context tokens. However, under few-shot supervision, even moderate label noise can disrupt prompt optimization. To address this issue, we propose TecoPrompt, a closed-loop robust prompt-learning framework that revisits optimal transport (OT) pseudo-labeling from a temporal perspective. TecoPrompt employs an entropic OT plan in the CLIP semantic space to obtain globally consistent label candidates. It verifies the reliability of these candidates by examining trajectory stability: a noisy label is only rewritten if the OT candidate remains unchanged within a K-epoch temporal stability window and passes a confidence gate based on Exponential Moving Average (EMA). This approach helps reduce confirmation bias. The rewritten labels are then integrated back into prompt training using a tri-group objective that includes three loss functions aligned with clean, mid, and noisy subsets. Experiments on seven datasets with synthetic symmetric and asymmetric noise, as well as Food101N, demonstrate significant performance improvements. For example, on the OxfordPets dataset, with 50% asymmetric noise, TecoPrompt achieves an accuracy of 0.843, up from 0.775.
Read the full abstract

All authors Zeyi Shao, Haowen Hua, Jiaxin Zhang, John See, Zeyd Boukhers, Cong Yang

Systems and Control

Two-stage Coordinated Energy Management of Train Operation and Wayside Energy Storage System for Rail Power Supply Systems

The increasing electrification of railway power supply system (RPSS) intensify the operational and economic challenges at the railway power system interface. Energy storage systems (ESSs) can provide fast and flexible support to mitigate short term power spikes and to improve the energy management. However, achieving coordinated operation is challenged by the tight coupling among electrical railway operation and ESS dispatch under time-varying traction demand and network limits. This paper proposes a two-stage coordinated energy management method for electrified RPSSs that jointly optimizes railway system operation, train trajectories and ESS dispatch while explicitly accounting for traction power flow constraints. First, a day-ahead operation stage determines the train operating profiles and the ESS setting decisions to establish the baseline operating plan. Then, an intra-day rolling optimization stage based on adaptive weight economic-model predictive control (AWC-MPC) updates ESS dispatch under refreshed forecasts of traction demand and renewable output. A real Swedish railway case is utilized to minimize the energy purchase cost and the ESS cost while reducing peak grid power, demonstrating its practical applicability with 35.7% peak grid power demand and 28.8% total system cost reduction.
Read the full abstract

All authors Fei Liu, Niklas Biedermann, Stefan Östlund, Qianwen Xu

Human-Computer Interaction

The Evolution of Coordination in a Collective Intelligence System: 25 Years of English Wikipedia and the Emergence of Generative AI

English Wikipedia is one of the largest examples of collective intelligence on the Web, sustained not only by article production but also by volunteer coordination and governance. While prior research has examined coordination work in Wikipedia, less attention has been paid to how participation in these spaces has evolved over time. Drawing on a longitudinal analysis spanning nearly 25 years of English Wikipedia, we examine editing patterns across five namespaces covering content, discussion, and governance. We find that participation in coordination spaces has declined relative to content production, particularly in governance areas, with a shrinking core of editors performing an increasing share of this work. Using Markov-based session metrics, we also find that editing has become more specialised, with editors moving less frequently between namespaces. Motivated by recent governance debates around generative AI, we conclude by investigating whether the availability of LLMs has altered these long-term trends. While short-term changes are visible, we find little evidence that generative AI fundamentally changed existing trajectories of coordination and participation.
Read the full abstract

All authors Neal Reeves, Maja Świeczkowska, Amy Rechkemmer, Elena Simperl

Audio and Speech Processing

Differentiable and Severity-invariant Discrete Tokens for Dysarthric Speech Recognition

This paper proposes novel differentiable and severity-invariant (DSI) discrete token approaches that are not only tightly integrated with downstream dysarthric speech recognition tasks, but also minimise discrete token diversity across speech impairment severity groups. Experiments conducted on the UASpeech and TORGO corpora suggest that Conformer models trained using the DSI tokens outperform the comparable baseline HuBERT discrete/continuous features by statistically significant WER reductions of 2.22%/0.78% absolute (9.14%/3.41% relative) and 1.78%/1.06% absolute (18.43%/11.86% relative) on the two tasks, respectively. After system combination, the lowest WERs of 18.90% and 6.38% were obtained on UASpeech and TORGO. Phoneme-specific T-SNE visualizations show that severity-invariant regularization reduces severity-dependent variation by producing greater overlap and less distinct boundaries among severity-group distributions.
Read the full abstract

All authors Huimeng Wang, Xurong Xie, Mengzhe Geng, Haoning Xu, Jiajun Deng, Youjun Chen, Chengxi Deng, Xunying Liu

Computation and Language

A Data-free Universal Prior over Syntactic Structures

The probabilities of syntactic structures in human languages are assumed to emerge fully from language-specific experience. Here, I show that a universal prior over syntactic structures emerges from a model of human language production, in which words are progressively integrated into syntactic structure. Without fitting any parameters to specific language data, the resulting prior assigns higher probabilities to attested than to random dependency trees in all 138 typologically diverse languages examined. These prior probabilities correlate positively with those estimated from corpora in 33 of 34 languages. The results indicate that part of the probability structure of syntax can arise independently of language-specific learning. This identifies human language production as a possible cognitive source of universal statistical structure in language, while providing a data-independent structural bias for probabilistic models, including large language models.
Read the full abstract

Machine Learning

Can Deep Learning Achieve Cross-Physics Mapping?

Can deep learning translate physical fields governed by fundamentally different equations? We address this question by introducing Cross-Physics Mapping (CPM), an operator-learning framework for mappings between heterogeneous physical domains. We formulate sufficient conditions for such mappings through compatible latent representations and propose a dimensionless scaling principle that aligns the characteristic evolution scales of the source and target systems without assuming their dynamical equivalence. As a representative test, paired diffusion and wave fields are generated independently from their respective parabolic and hyperbolic equations while sharing the same latent geometry, material heterogeneity, excitation, and dimensionless scale. Seven architectures-ResUNet, DeepONet, Fourier, latent, wavelet, U-shaped, and Galerkin neural operators-are evaluated for both diffusion-to-wave and wave-to-diffusion mappings. The results reveal a strong directional asymmetry. Diffusion-to-wave reconstruction is more challenging because it requires recovering wavefront, phase, and time-of-flight information attenuated by diffusion; U-NO performs best in this direction, achieving a relative $\ell_2$ error of $0.307$ and an $R^2$ of $0.905$. Wave-to-diffusion mapping is considerably more stable, with GNO attaining a relative $\ell_2$ error of $0.154$ and an $R^2$ of $0.935$. Neural operators generally outperform the conventional convolutional baseline, highlighting the nonlocal nature of cross-physics transformations. These findings demonstrate that deep learning can establish useful mappings between distinct physical modalities on a shared latent manifold, while the achievable accuracy remains fundamentally constrained by the direction-dependent information content of the governing physics.
Read the full abstract

Artificial Intelligence

CoAdapt: An LLM-based Framework for Adaptive Collaborative Perception in IIoT Robotic Swarms

Industrial IoT environments increasingly deploy autonomous mobile robots for tasks such as material handling, product assembly, or infrastructure inspection. In such deployments, collaborative perception enables robots to share LiDAR observations and collectively construct a richer model of their environment than an individual agent could produce alone. However, industrial environments are dynamic spaces where robot positions shift continuously, network bandwidth fluctuates, and the marginal contribution of robots to perception quality varies at runtime. Existing collaborative perception approaches are designed for static participation assumptions and cannot adapt to these dynamics without sacrificing either detection precision or communication efficiency. This paper presents CoAdapt, an adaptive collaborative perception framework for IIoT robotic swarms in which a Large Language Model (LLM) serves as a runtime fusion controller, jointly deciding which robots participate in the fusion process and which fusion algorithm to apply based on the current spatial configuration and network state. The LLM reasons over structured natural language descriptions of the scene derived from raw LiDAR point clouds, requiring no taskspecific training and generalizing to unseen swarm topologies. Evaluated on the OPV2V benchmark across 25 scenarios, our approach achieves a 38% reduction in communication cost while maintaining detection precision comparable to static baseline approaches.
Read the full abstract

All authors Houssam Hajj Hassan, Antonia Maria Masucci, Lynda Zitoune, Salah-Eddine Elayoubi

Quantum Physics

Understanding the spin coherence of molecular photoexcited triplet states from first principles

Optically readable molecular spins are attractive as quantum sensors due to their nanoscale modularity, synthetic tunability, and scope for sensitive readout. In particular, photoexcited spin-triplet states in organic molecules are appealing in supporting high optical-spin contrast at room temperature. Their utility is underpinned by their coherence, which warrants a detailed understanding of it. Here, from first principles, we systematically explore the Hahn-echo decoherence of the benchmark molecular system for room-temperature optically detected spin coherence—pentacene guest molecules coupled to a para-terphenyl host. Using generalized cluster-correlation expansion methods, we investigate the mechanisms of nuclear-spin-induced decoherence from zero to high magnetic field, exploring the role of guest vs host molecules, specific nuclei, zero-field splitting interactions, and hyperfine parameters. We describe how zero-field decoherence is driven by ~6 nuclei on the guest, while high-field decoherence is driven by ~600 nuclei in the host; how the longitudinal zero-field splitting parameter, $D$, can prolong $T_2$; and the magnetic-field-dependent processes which drive decoherence. These results advance our understanding of decoherence in optically readable molecular spins, providing insight for their synthetic enhancement and deployment as quantum probes.
Read the full abstract

Information Retrieval

Efficient Swing Computation for Retrieval in Large-Scale Recommender Systems

Given a user-item graph $G$, a query item $v_q$ and a target item $v_t$, the Swing score $sw(v_q, v_t)$ of the item pair $(v_q, v_t)$ leverages the user-item-user interaction structure to evaluate their similarity. This measure is found to be highly effective in item-to-item (i2i) retrieval task and finds extensive applications in industrial-scale recommender systems. However, existing solutions towards computing Swing scores are either prohibitively expensive due to their quadratic time complexity w.r.t. the item degree, or rely on truncation heuristics that yield unsatisfactory quality, rendering them impractical particularly on graphs with billions of interactions. In this paper, we present ASC and $K$-ASC, two novel and efficient algorithms for approximate and top-$K$ Swing queries, to address the aforementioned limitations. Specifically, these algorithms provide rigorous theoretical guarantees in probabilistic relative and additive errors of Swing values. The basic idea of ASC is to combine two randomized algorithms, GNS and USS, in a simple yet non-trivial way to adaptively process high- and low-degree query items with minimal runtime cost. In particular, $K$-ASC offers practical efficiency and effectiveness for top-$K$ queries through a filter-refinement paradigm with carefully-designed heuristics. Extensive experiments over eight real datasets demonstrate that ASC and $K$-ASC can achieve orders of magnitude speed-up over competitors in terms of computational time while offering the same approximate and top-$K$ query result quality, and in particular, $K$-ASC is highly efficient on massive graphs including the billion-edge Yambda and MAG datasets.
Read the full abstract

Combinatorics

Hall algebras of graphs and rooted trees

Hall algebras can be associated with a broad class of combinatorial structures through the theory of 2-Segal sets. In this paper, we study the Hall algebras arising from the 2-Segal sets of rooted trees, undirected graphs, and directed graphs. In each case, we establish an analogue of Green's theorem giving a twisted bialgebra structure to the Hall algebra, describe the primitive elements, and derive a presentation by generators and relations. As an application, we realize the Hall algebra of an undirected graph as the cohomology ring of a topological space. In the case of directed graphs, we introduce the Hall polynomial, defined as the Poincaré polynomial of the space of primitive elements. The Hall polynomial is an invariant of the underlying undirected graph which we show to be closely related to the Tutte polynomial. We furthermore give examples of graphs with equal Tutte polynomials but distinct Hall polynomials showing thus that the Hall polynomial encodes different information from that of the Tutte polynomial.
Read the full abstract

Group Theory

The Fourth Continuous Bounded Cohomology of the Complex Symplectic Group

We prove that $H_{\mathrm{cb}}^4(\Sp(4,\CC);\RR)=0$. Together with Blatz's secondary stability and the rank-one theorem of Bucher–Savini, this gives degree-four vanishing for all complex symplectic and odd complex orthogonal groups. In normalized symplectic Gram coordinates, we establish a bounded-primitive estimate and compute the measurable cohomology of the projective action. An explicit rational cocycle has divergent periods on a family of finite orbit cycles of uniformly bounded $\ell^1$-mass. A two-cone averaging construction extends the period estimate to measurable cochains and excludes bounded representatives of every nonzero degree-four measurable action class.
Read the full abstract

Computer Vision

RegRet: Enhancing Region-Level Retrieval in Large Multimodal Models

Region-level retrieval aims to align user-specified image regions with relevant regions or textual descriptions, playing a crucial role in realworld applications such as e-commerce product search and RAG. Although recent Large Multimodal Models (LMMs) have made significant strides in multimodal retrieval, they primarily focus on global-level tasks and struggle to capture effective region-level representations. To bridge this gap, we present RegRet, an LMM-based Region-level Retrieval framework that enhances the regional representations without compromising overall global retrieval performance. At its core, RegRet integrates a Region-Aware Encoder to capture detailed regional features while balancing them with the global background context. To further enhance the fine-grained understanding and discriminability of representations, we design a multi-stage training pipeline that includes detailed localized captioning and regional contrastive learning tasks. In addition, considering the absence of region-level contrastive training data and the limited diversity of evaluation tasks in current benchmarks, we introduce the REGMB benchmark. It comprises 225k contrastive pairs, covering four multimodal retrieval tasks. Extensive experiments validate the effectiveness of our approach. RegRet outperforms strong baselines in the zero-shot setting. Further training with contrastive learning leads to an average improvement of more than 20% on both REGMB and public benchmarks, while achieving comparable or better results on global-level retrieval tasks.
Read the full abstract

All authors Xun Liang, Honghui Yang, Weihang Pan, Ruisi Zhao, Boyuan Pan, Yao Hu, Wenxiao Wang, Binbin Lin, Deng Cai

Neural and Evolutionary Computing

LLMDE: A Large Language Model-Driven Differential Evolution Algorithm for Portfolio Optimization

This study proposes a Large Language Model-Driven Differential Evolution (LLMDE) algorithm to reduce the reliance on handcrafted hyperparameter design. The proposed algorithm leverages a prompt engineering strategy, allowing large language models (LLMs) to dynamically select mutation strategies and configure control parameters guided by optimization feedback, thus enhancing the performance of the DE algorithm. We evaluate the performance of LLMDE on the CEC2022 benchmark suite, comparing it with standard DE and representative metaheuristics. Furthermore, we employ factor analysis and K-means clustering for stock selection, and then apply LLMDE to solve the Conditional Value at Risk (CVaR) portfolio optimization problem using the selected stocks, subject to budget and minimum expected return constraints. Experimental results demonstrate that LLMDE achieves competitive performance on the benchmark suite while continuously generating high-quality solutions for complex constrained optimization tasks. These outcomes successfully demonstrate the viability of embedding LLMs within metaheuristics, paving a promising path toward the design of advanced LLM-assisted optimization techniques.
Read the full abstract

All authors Rong Chai, Vaclav Snasel, Xiaopeng Wang, Seyedali Mirjalili, Crina Grosan

Physics Education

Analytical and Experimental Study of a Variable-Mass Oscillator with Constant Mass Loss

We present an analytical and experimental study of a variable-mass oscillator with constant mass loss. Starting from the equation of motion of a spring-mass system whose mass decreases linearly with time, analytical solutions are obtained in terms of Bessel functions. Simple asymptotic expressions are then derived, providing explicit descriptions of the evolution of the oscillation amplitude and frequency. The analysis is extended to include viscous damping, leading to a model suitable for direct comparison with experimental observations. Experimental measurements of the oscillator acceleration were performed using a simple laboratory setup consisting of a spring-mounted container that continuously loses mass at an approximately constant rate. The theoretical predictions accurately reproduce the measured dynamics over the entire duration of the oscillation, and the massloss rates obtained from the fits are in good agreement with independent measurements. The results provide an accessible example of variable-mass dynamics and illustrate how special-function methods can yield experimentally testable predictions in advanced undergraduate mechanics.
Read the full abstract

All authors Rod Milbrandt, Josep Ll. Suñer, Juan C. Castro-Palacio, Francisco M. Muñoz-Peréz, Juan A. Monsoriu

General Relativity and Quantum Cosmology

An Abel-Inversion Formalism for Spacetime Metric Reconstruction from Light Deflection

We develop an inverse lensing formalism for reconstructing the metric function of a static, spherically symmetric spacetime directly from the gravitational deflection angle of light. By formulating the inverse problem through an Abel transformation, we derive an integro-differential relation connecting the observable deflection profile to the underlying spacetime geometry. As a consistency check, we consider the case of vanishing deflection, \(\alpha(b)=0\), and recover \(A(r)=1\), corresponding to flat Minkowski spacetime. We then apply the formalism to the gravitational deflection of light by the Sun using the observationally motivated leading-order expression $\alpha(b)=\frac{2(1+\gamma)GM_{\odot}}{c^{2}b}.$The resulting metric function is shown to recover the Schwarzschild form in the weak-field limit \(r\gg M\). We further consider the higher-order correction to the solar deflection angle and reconstruct the corresponding metric beyond the leading-order approximation. Interestingly, the \((M/r)^2\) term vanishes in the resulting weak-field expansion, yielding an improved approximation compared with the metric reconstructed from the leading-order deflection angle. Our results demonstrate that gravitational lensing observations can provide a direct route to reconstructing the underlying spacetime geometry without assuming a specific metric a priori. This formalism therefore, offers a novel framework for connecting observational light propagation with spacetime geometry.
Read the full abstract

Number Theory

On $t$-sumfree sequences

In this paper we explore the world of $t$-sumfree sequences, primarily for $t=2$ and $t=3$. Starting with $t$ increasing positive integers, build an infinite sequence by choosing as the next entry the least positive integer that is not the sum of $t$ distinct previous values. The main open problem concerns the question of the conjectured ultimate periodicity of these sequences. We formulate a sufficient condition to prove that a conjectured preperiod and period are correct. We also prove periodicity for a considerable subclass of 2-sumfree and of 3-sumfree sequences. Finally, we present the results of extensive computations on 3-sumfree sequences.
Read the full abstract

Computer Vision

FAHCD-Net: Frequency-Adaptive Heatmap-Conditional Diffusion Networks for Robust Facial Landmark Detection

Facial Landmark Detection(FLD) is a crucial task in various applications and has achieved significant advancements in recent years. However, current FLD methods still struggle under challenging conditions, where facial structural variations, information loss, and noise interference severely compromise the integrity and accuracy of learned facial features. To address these issues, we propose Frequency-Adaptive Heatmap-Conditional Diffusion Network (FAHCD-Net), which integrates a Frequency-Adaptive Heatmap-Conditional Diffusion (FAHCD) model with a Smoothness Regularization (SR) loss in a cascaded framework. Specifically, the FAHCD model incorporates a Hierarchical Frequency Adaptation (HFA) module designed to suppress redundant high-frequency noise through multi-layer frequency decomposition and adaptive reconstruction, thereby preserving essential facial structures. Additionally, the SR loss is proposed to further mitigate the interference of high-frequency noise and enhance the smoothness of the generated landmark heatmaps. By cascading the FAHCD model with the SR loss, FAHCD-Net effectively leverages both statistical and frequency-based distribution characteristics of the data to progressively generate more accurate landmark heatmaps from noisy inputs. Extensive experiments on popular benchmarks demonstrate the effectiveness and robustness of the proposed method, achieving state-of-the-art performance in FLD tasks under challenging scenarios. The source code is available at https://github.com/HJWKryptonite/FAHCD-Net.
Read the full abstract

All authors Jun Wan, Jiwei Hu, Shengkai Hu, Qilu Zhu

Computer Vision

StackTok: Accelerating VLMs Inference with Budget-Adaptive Visual Token Selection

Increasing image resolution produces ever-longer visual-token sequences in vision-language models (VLMs), substantially raising their inference cost. To reduce this overhead without retraining, existing methods select compact token subsets that prioritize query relevance, visual coverage, or a fixed trade-off between them. The appropriate balance, however, varies across queries and token budgets: localized questions favor relevance, whereas holistic questions demand broader visual coverage. We introduce StackTok, a training-free selector that treats query relevance as the objective and visual coverage as budget-calibrated support. StackTok builds a size-indexed coverage reference from a coverage-only greedy sequence and adjusts its support target using query–vision affinity entropy. A reference-gated interleaved selection policy then switches between relevance- and coverage-oriented additions according to the current subset's support deficit. For high-resolution inputs, StackTok allocates one shared token budget across crops according to the combined marginal gain of locally nominated tokens. Evaluated with five VLMs over ten distinct image-understanding benchmarks, StackTok ranks first among training-free selectors in every tested model–budget setting. On high-resolution LLaVA-NeXT-7B, it retains 95.26% of full-token performance with only 160 of 2{,}880 (5.6%) visual tokens.
Read the full abstract

All authors Zhenbin Wang, Lei Zhang, Lituan Wang, Wei Huang, Yan Wang, Zhenwei Zhang

Geometric Topology

Splice formulae for Poincaré series of integral homology spheres

Let M be a plumbed integral homology sphere 3-manifold associated with a connected negative definite plumbing graph $\Gamma$. One of its most important invariants is its multivariable Poincaré series (or zeta function) $f_\Gamma(\mathbf {t})$. Several numerical invariants can be read from $f_\Gamma(\mathbf{t})$, or even from its `polynomial part' $ \operatorname{Pol}_{\Gamma}(\mathbf{t})$. For example, the `normalized' Casson's invariant equals $ \operatorname{Pol}_{\Gamma}(1)$. In this note we provide several splice (surgery) formulae for $f_\Gamma$ and $\operatorname{Pol}_\Gamma$ (reduced to the node variables, or to the node variables of connected sub-graphs). In this way, these global invariants can be recovered from a collection of certain smaller graphs. Recall that some (integral homology sphere) 3-manifold invariants are additive with respect to the splice decomposition (like the Casson's invariant). However, some invariants need some `splice correction terms'. In our formulae the correction terms are easily computable one variable Alexander polynomials.
Read the full abstract

Networking

The Price of the Golden 6G Band: Evaluation of Beam Management Effort in FR3

Frequency Range 3 (FR3), 7.125-24.25 GHz, regarded as the "golden band" for 6G networks, has less challenging propagation characteristics than FR2 while offering much wider bandwidth for high data rate applications than FR1. Reusing existing FR1 infrastructure for FR3 network deployments requires gNodeBs (gNBs) to employ antenna arrays and perform beam management, which has proven challenging at FR2. In this paper, we extensively study and characterize the beam management effort in an FR3 urban network, in terms of: beam alignment sensitivity, number of directional link opportunities, gNB handover and beam switch rates, and beam steering distance. Our results show that achieving a high and stable mobile throughput requires significant beam management effort across FR3 bands. While the beam tracking requirements are less stringent at the lower frequencies due to wider beams, the beam switching rate to a non-adjacent beam is relatively comparable at FR3 and FR2.
Read the full abstract

Fluid Dynamics

Direct Trajectory Reconstruction for Fast 3D X-ray Particle Tracking Velocimetry in Porous Media

X-ray tomographic particle tracking velocimetry is a promising tool to provide pore-scale insight into fluid flow inside opaque porous media. This is critical for applications ranging from geo-energy to electrochemistry. Current approaches to implement this three-dimensional, three-component (3D-3C) velocimetry technique use conventional micro-computed tomography (CT) workflows to recover tracer particle trajectories. These workflows first reconstruct a sequence of 3D CT images from radiographs which are sequentially acquired at time intervals that span multiple viewing angles. The tracer particles are then detected in each of these CT images and finally linked into trajectories. However, motion-blurring induced by the static CT volume reconstruction impairs the detectability of fast-moving tracer particles, severely limiting the temporal resolution and dynamic range of the velocimetry outcome. In this work, we show that this limitation can be overcome by explicitly recovering particle trajectories directly from the acquired radiographs, bypassing the intermediate tomographic reconstruction step and the static-scene assumptions of traditional 3D CT reconstructions. We implement this approach in a new open-source algorithm, CTracks, and validate it using experimental data on pipe flow and on flow through a complex porous medium, demonstrating a five-fold increase in the maximum detectable velocity compared with state-of-the-art methods. Numerical simulations further confirm improved detection rates and bounded trajectory errors at these increased velocities. Our work thus demonstrates that 3D-3C velocimetry in opaque microscopic geometries can be extended to substantially faster flows by replacing X-ray tomographic image reconstruction with direct particle trajectory reconstruction.
Read the full abstract

All authors Robert van der Merwe, Wannes Goethals, Sharon Ellman, Sojwal Manoorkar, Jan Aelterman, Matthieu N. Boone, Tom Bultreys

Information Theory

On Sequence Reconstruction Problem for q-ary Deletion Channels

The sequence reconstruction problem for $q$-ary deletion channels, introduced by Levenshtein in 2001, concerns the minimum number of channels required to uniquely recover a transmitted sequence when each channel introduces exactly $t$ deletions. Combinatorially, it is equivalent to determining $N_q(n,d,t)$, the maximum intersection size of two $t$-deletion balls with centers at Levenshtein distance at least $d$, for $q$-ary sequences of length $n$ over the alphabet \(\Sigma_q=\{0,1,\dots,q-1\}\). Levenshtein solved the uncoded case $N_q(n,1,t)$ for all $n\ge t$; subsequently, Gabrys and Yaakobi determined $N_2(n,2,t)$, and Wang et al. extended the result to $N_3(n,2,t)$. In this paper, we study the problem for \(q\)-ary sequences under minimum Levenshtein distance \(d=2\) with channels that introduce exactly \(t\) deletions. We determine the exact value of \(N_q(n,2,t)\) for all \(t\ge 2, q\geq 4\), and for sufficiently large \(n\), and construct explicit pairs of sequences attaining the maximum intersection. Furthermore, for each $q\ge3$, we characterize all extremal sequence pairs. In particular, if the intersection size matches the first two terms of \(N_q(n,2,t)\), then the two center sequences must contain, at the same positions, length-5 blocks of the forms \((a,b,c,a,b)\) and \((b,a,c,b,a)\) for some distinct \(a,b,c\in\Sigma_q\); for \(t\ge q+2\), the exact maximum \(N_q(n,2,t)\) is attained precisely by \(2q!\) unordered pairs of sequences with a specific block structure. Asymptotically, we prove that for \(q\ge 4\) and \(t\ge 2\), \[ N_q(n,2,t)=\frac{6}{(t-2)!}n^{t-2}-\frac{3t+13}{(t-3)!}n^{t-3}+\frac{3t^2+25t+64}{4(t-4)!}n^{t-4}+O(n^{t-5}). \] Moreover, \(N_q(n,2,t)\) and \(N_{q-1}(n,2,t)\) share their first \(q-1\) terms, and for \(t\ge q\) the coefficient of \(n^{t-q}\) in their difference is \(\frac{6t-6q+5}{(t-q)!}\).
Read the full abstract

Information Theory

CPM-LDPC Codes Attaining the Minimum-Distance Bound

We study binary quasi-cyclic LDPC codes whose parity-check matrices are full arrays of single circulant permutation matrices (CPMs), referred to here as CPM-LDPC codes. Their minimum distance is at most $(J+1)!$, where $J$ is the column weight. For every fixed pair of column and row weights $2\le J<L$, we show that this bound is attained for all sufficiently large integer lift sizes. First, we give one integer exponent matrix independent of the lift size $P$. Second, we show that independent uniform exponent choices attain the bound with probability $1-O_{J,L}(P^{-1})$. Both proofs use cycle conditions required by low-weight codewords and a lower bound on the number of terms in vectors satisfying polynomial check equations. Neither construction requires $P$ to be prime. We also give small-lift arrays attaining the bound 24 for $J=3$, $L=4,\ldots,8$, and arrays with distance at least 28 for $J=4$, $L=5,\ldots,8$, together with computational distance verification.
Read the full abstract