October 2025 arXiv papers — page 152

David Benisty, Antonino Del Popolo

The local Hubble flow provides a valuable probe of the transition between cosmic expansion and nonlinear gravitational dynamics. On large scales, galaxies follow the linear Hubble law, but within group- and cluster-sized environments, gravitational interactions generate substantial deviations. Using the IllustrisTNG cosmological simulations, we test whether

Othman Cherkaoui Dekkaki

We develop an optimal control model for allocating agricultural crop residues between bioenergy production and soil fertility restoration. The system captures a novel circular feedback: a fraction of cumulative energy output is reinvested into soil productivity, linking energy use with ecological regeneration. The dynamics are governed by a nonlinear three-s

Abdullah Al Mahmud, Prangon Chowdhury, Mohammed Borhan Uddin, Khaled Eabne Delowar

Anemia, a condition marked by insufficient levels of red blood cells or hemoglobin, remains a widespread health issue affecting millions of individuals globally. Accurate and timely diagnosis is essential for effective management and treatment of anemia. In recent years, there has been a growing interest in the use of artificial intelligence techniques, i.e.

Tianshun Han, Benjia Zhou, Ajian Liu, Yanyan Liang

PESTalk is a novel method for generating 3D facial animations with personalized emotional styles directly from speech. It overcomes key limitations of existing approaches by introducing a Dual-Stream Emotion Extractor (DSEE) that captures both time and frequency-domain audio features for fine-grained emotion analysis, and an Emotional Style Modeling Module (

Tatsuro Kawakami, Jakub Witaszek

We prove inversion of adjunction for higher rational singularities.

Kotaro Motegi

We prove that if a one-parameter family of varifolds has an $L^2$ normal velocity $v$ in the sense of Brakke, and if the family is represented as the graph of a continuous function $f$ with continuous spatial derivative $\nabla f$, then $f$ has weak derivatives $\partial_t f, \nabla^2 f \in L^2$, and $v$ coincides with the usual normal velocity of the graph.

Guoqing Tian, Li-Li Zheng, Zhi-Ming Zhan, Franco Nori

Strongly correlated photons play a crucial role in modern quantum technologies. Here, we investigate the probability of generating strongly correlated photons in a chain of N qubits coupled to a one-dimensional (1D) waveguide. We found that disorder in the transition frequencies can induce photon antibunching, and especially nearly perfect photon blockade ev

Nadezhda Zolotova, Mikhail Vokhmyanin

In this work, we present an extensive review and detailed analysis of sunspot measurements, drawings, and engravings made by John Flamsteed and, mainly, by Philippe de La Hire during the Maunder minimum. All available information and contemporary knowledge about the sunspot nature are shown. The coordinates, areas, and numbers of sunspots and sunspot groups

Ruiqi Kong, He Chen

WiFi sensing based on channel state information (CSI) collected from commodity WiFi devices has shown great potential across a wide range of applications, including vital sign monitoring and indoor localization. Existing WiFi sensing approaches typically estimate motion information directly from CSI. However, they often overlook the inherent advantages of ch

Dean L. Slack, Noura Al Moubayed

Although large language models excel across many tasks, they can memorise training data and thereby expose private or copyrighted text. Most defences target the pre-training stage, leaving memorisation during fine-tuning, especially for domain adaptation and instruction tuning, poorly understood. We fine-tune Pythia, Llama3, and Mistral models spanning 1.4B-

Jens Marklof, Andreas Strömbergsson, Shucheng Yu

This paper extends a recent extreme value law for horocycle flows on the space of two-dimensional lattices, due to Kirsebom and Mallahi-Karai, to the simplest examples of rank-$k$ unipotent actions on the space of $n$-dimensional lattices. We analyse the problem in terms of the hitting time and impact statistics for the unipotent action with respect to a shr

Wenhan Ma, Hailin Zhang, Liang Zhao, Yifan Song

Reinforcement learning (RL) has emerged as a crucial approach for enhancing the capabilities of large language models. However, in Mixture-of-Experts (MoE) models, the routing mechanism often introduces instability, even leading to catastrophic RL training collapse. We analyze the training-inference consistency of MoE models and identify a notable discrepanc

Shijie Zhao, Xuanyu Zhang, Weiqi Li, Junlin Li

Reasoning-based image quality assessment (IQA) models trained through reinforcement learning (RL) exhibit exceptional generalization, yet the underlying mechanisms and critical factors driving this capability remain underexplored in current research. Moreover, despite their superior performance, these models incur inference energy usage and latency orders of

Eric Jeangirard

We present a dataset of 833k paragraphs extracted from CC-BY licensed scientific publications, classified into four categories: acknowledgments, data mentions, software/code mentions, and clinical trial mentions. The paragraphs are primarily in English and French, with additional European languages represented. Each paragraph is annotated with language ident

Kleitos Papadopoulos

This paper addresses the single-item lot sizing problem with a 1-breakpoint all-units quantity discount in a monotonic setting where the purchase prices are non-increasing over the planning horizon. For this case, we establish several novel properties of the optimal solution and develop a hybrid dynamic programming approach that maintains a compact represent

S. Tarasov

We compute generating functions of the set of directed lattice paths starting from the origin and avoiding a periodic set of even point on OX = "time"-axis. As an application we prove a combinatorial identity proposed by P. Hajnal and G.V. Nagy.

Ruben Johnson Robert Jeremiah, Peyman Goli, Steven van de Par

Separating competing speech in reverberant environments requires models that preserve spatial cues while maintaining separation efficiency. We present a Phase-aware Ear-conditioned speaker Separation network using eight microphones (PEASE-8) that consumes complex STFTs and directly introduces a raw-STFT input to the early decoder layer, bypassing the entire

Kon Woo Kim, Rezarta Islamaj, Jin-Dong Kim, Florian Boudin

This study investigates how existing annotation guidelines can be repurposed to instruct large language model (LLM) annotators for text annotation tasks. Traditional guidelines are written for human annotators who internalize training, while LLMs require explicit, structured instructions. We propose a moderation-oriented guideline repurposing method that tra

Xuwang Yin, Claire Zhang, Julie Steele, Nir Shavit

Simultaneously achieving robust classification and high-fidelity generative modeling within a single framework presents a significant challenge. Hybrid approaches, such as Joint Energy-Based Models (JEM), interpret classifiers as EBMs but are often limited by the instability and poor sample quality inherent in training based on Stochastic Gradient Langevin D

Vedant Dhruv, Ben Prather, Mani Chandra, Abhishek V. Joshi

The black holes in the Event Horizon Telescope sources Messier 87* and Sagittarius A* (SgrA*) are embedded in a hot, collisionless plasma that is fully described in kinetic theory yet is usually modeled as an ideal, magnetized fluid. In this Letter, we present results from a new set of weakly collisional fluid simulations in which leading order kinetic effec

Anjan Kar, Ayan Dey, Sayan Kar

We investigate the Penrose process of energy extraction in the context of rotating regular black holes. For the Neves-Saa class of regular black hole solutions, which includes the Bardeen, Hayward and Fan-Wang spacetimes as special cases, the extraction efficiency is bounded above by the known value for Kerr spacetime. However, in the case of a new rotating

Tian-Nuo Li, Guo-Hong Du, Yun-He Li, Yichao Li

Recent DESI baryon acoustic oscillation (BAO) measurements, combined with Planck cosmic microwave background (CMB) data and DESY5 type Ia supernova (SN) data, indicate a significant deviation from $Λ$CDM, which seems to suggest that this deviation can be explained by an interaction between dark energy and dark matter. In this work, we perform a comprehensive

Sayan Sarkar, Sunit Das, Amit Agarwal

We demonstrate a deterministic switching mechanism in collinear altermagnets driven by asymmetric sublattice spin currents. Unlike conventional antiferromagnets, where combined parity-time-reversal symmetry enforces purely staggered sublattice spin torques, altermagnets host symmetry-protected nonrelativistic spin splitting that produces unequal torques on t

Leandro Soares Indrusiak

We present a novel protocol that reduces worst-case packet latency in deflection-based on-chip interconnect networks. It enforces the deflection of the header of a packet but not its payload, resulting in a reduction in overall network traffic and, more importantly, worst-case packet latency due to decreased pre-injection latency.

Milon Bhattacharya

Quick commerce (q-commerce) is one of the fastest growing sectors in India. It provides informal employment to approximately 4,50,000 workers, and it is estimated to become a USD 200 Billion industry by 2026. A significant portion of this industry deals with perishable goods. (e.g. milk, dosa batter etc.) These are food items which are consumed relatively fr

Laurine Martinien, Gaspard Duchêne, François Ménard, Karl R. Stapelfeldt

The James Webb Space Telescope provides unprecedented information to study ices in protoplanetary disks. However, the saturation of ice bands in highly inclined disks hinders the measurement of ice abundances using classical spectroscopy. This is unfortunate as the presence and more importantly abundance of ices plays a key role in, e.g., the evolution of du

Hengran Zhang, Keping Bi, Jiafeng Guo, Jiaming Zhang

Retrieval-augmented generation (RAG) is typically optimized for topical relevance, yet its success ultimately depends on whether retrieved passages are useful for a large language model (LLM) to generate correct and complete answers. We argue that such utility is often LLM-specific rather than universal, due to differences in models' knowledge, reasoning

José Geraldo Telles Ribeiro, Americo Cunha

Polymer-based plastics exhibit time-dependent deformation under constant stress, known as creep, which can lead to rupture or static fatigue. A common misconception is that materials under tolerable static loads remain unaffected over time. Accurate long-term deformation predictions require experimental creep data, but conventional models based on simple rhe

Edmundo J. Huertas, Alberto Lastra, Judit Minguez Ceniceros

The work analyzes the theory of Dunkl operator as a moment differential operator. This last operator generalizes the first one whenever the sequence of moments satisfies appropriate classical properties, classically considered in the general theory of ultraholomorphic and ultradifferentiable classes of functions. In this sense, the theory of Dunkl operator i

Nixon Hanna, Paul L. Schechter, Michael A. McDonald, Marceau Limousin

For any elliptical potential with an external parallel shear, Witt has proven that the gravitational center lies on a rectangular hyperbola derived from the image positions of a single quadruply lensed object. Moreover, it is predicted that for an isothermal elliptical potential the source position both lies on Witt's Hyperbola and coincides with the center

Tim Kokkeler, Tero T. Heikkilä, F. Sebastian Bergeret

We study the nonequilibrium spin-splitter effect in superconducting altermagnets and superconductor altermagnet hybrids by computing the alternating spin current and edge the spin density in the presence of an alternating electric field. We show that while in the normal state the effect is not sensitive to the field frequency, in the superconducting state, t

Xuan Tang, Han Zhang, Yuan Cao, Difan Zou

Adam is a popular and widely used adaptive gradient method in deep learning, which has also received tremendous focus in theoretical research. However, most existing theoretical work primarily analyzes its full-batch version, which differs fundamentally from the stochastic variant used in practice. Unlike SGD, stochastic Adam does not converge to its full-ba

Woo-Hyun Ko, Jaewon Kim, Tzu-Hsiang Lin, Samin Moosavi

We consider a problem faced by an intelligent roadside unit (RSU) monitoring a roadway by a video camera. Suppose the RSU notices that a particular car in its visual field needs to execute a specific evasive maneuver to avoid danger. It would like to send a packet addressed to that particular car with this suggestion. The problem is that while all the cars a

M. P. Snelders, J. W. T. Hessels, J. Huang, N. Sridhar

FRB 20121102A is the original repeating fast radio burst (FRB) source and also the first to be localised to milliarcsecond precision using very-long-baseline interferometry (VLBI). It has been active for over 13 years and resides in an extreme magneto-ionic environment in a dwarf host galaxy at a distance of ~1 Gpc. In this work, we use the European VLBI Net

Lucie E. Rowland, Kasper E. Heintz, Hiddo Algera, Mauro Stefanon

Neutral gas in galaxies during the Epoch of Reionisation regulates star formation, dust growth, and the escape of ionising photons, making it a key ingredient in understanding both galaxy assembly and reionisation. Yet, direct constraints on the HI content of galaxies at z>6 have been scarce. With JWST, Ly$\alpha$ damping wings in galaxy spectra can now prov

Andrea Di Biagio, Carlo Rovelli

We offer a fresh perspective on the relational interpretation of quantum mechanics as a way of thinking about the world described by quantum theory based on quantifiable notions of information. This allows us to provide a definition of a relative fact, with no addition to orthodox quantum theory and no fundamentally special role for observers. By associating

Elwin Huaman, Wendi Huaman, Jorge Luis Huaman, Ninfa Quispe

Under-resourced languages, such as Quechuas, face data and resource scarcity, hindering their development in speech technology. To address this issue, Common Voice presents a crucial opportunity to foster an open and community-driven speech dataset creation. This paper examines the integration of Quechua languages into Common Voice. We detail the current 17

Patrick Bastian, Tim Kutta

We propose a new class of sequential change point tests, both for changes in the mean parameter and in the overall distribution function. The methodology builds on a two-window inspection scheme (TWIN), which aggregates data into symmetric samples and applies strong weighting to enhance statistical performance. The detector yields logarithmic rather than pol

Etzion Harari, Moshe Unger

Graph Neural Networks (GNNs) have demonstrated remarkable success in node classification tasks over relational data, yet their effectiveness often depends on the availability of complete node features. In many real-world scenarios, however, feature matrices are highly sparse or contain sensitive information, leading to degraded performance and increased priv

Joshua Niemeijer, Jan Ehrhardt, Heinz Handels, Hristina Uzunova

Generative Models are a valuable tool for the controlled creation of high-quality image data. Controlled diffusion models like the ControlNet have allowed the creation of labeled distributions. Such synthetic datasets can augment the original training distribution when discriminative models, like semantic segmentation, are trained. However, this augmentation

Han Lu, Zichen Liu, Shaopan Xiong, Yancheng He

Synchronous Reinforcement Learning (RL) post-training has emerged as a crucial step for enhancing Large Language Models (LLMs) with diverse capabilities. However, many systems designed to accelerate RL post-training still suffer from low resource utilization and limited scalability. We present ROLL Flash, a system that extends ROLL with native support for as

Jason Veara, Manav Jain, Kyle Moy, Aanjhan Ranganathan

Mysterious sightings of Unmanned Aircraft Systems (UAS) over U.S. military facilities, suburban neighborhoods, and commercial airports have intensified scrutiny of drone activity. To increase accountability, the Federal Aviation Administration (FAA) introduced a Remote ID mandate, requiring unmanned aircraft to broadcast their location, operator's location,

Max Tymczyszyn, Edward McCann

We describe nonsymmorphic four-band tight-binding models in one dimension with Kramers degeneracy, and propose topolectric-circuit realizations of their topological phases. We begin with a representative model in the nonsymmorphic AII class with symmorphic time-reversal symmetry and nonsymmorphic charge-conjugation and chiral symmetries, resulting in a $\mat

Haomin Wang, Jinhui Yin, Qi Wei, Wenguang Zeng

General SVG modeling remains challenging due to fragmented datasets, limited transferability of methods across tasks, and the difficulty of handling structural complexity. In response, we leverage the strong transfer and generalization capabilities of multimodal large language models (MLLMs) to achieve unified modeling for SVG understanding, editing, and gen

Zhao Huang, Boyang Sun, Alexandros Delitzas, Jiaqi Chen

Interactive 3D scenes are increasingly vital for embodied intelligence, yet existing datasets remain limited due to the labor-intensive process of annotating part segmentation, kinematic types, and motion trajectories. We present REACT3D, a scalable zero-shot framework that converts static 3D scenes into simulation-ready interactive replicas with consistent

Hui-Li Han, Chen Wang

In this paper, we study some supercongruences involving the sequence $$ t_n(x)=\sum_{k=0}^n\binom{n}{k}\binom{x}{k}\binom{x+k}{k}2^k $$ and solve some open problems. For any odd prime $p$ and $p$-adic integer $x$, we determine $\sum_{n=0}^{p-1}t_n(x)^2$ and $\sum_{n=0}^{p-1}(n+1)t_n(x)^2$ modulo $p^2$; for example, we establish that \begin{align*} \sum_{n=0}

K. E. Feldman

A new semi-analytical pricing model for Bermudan swaptions based on swap rates distributions and correlations between them. The model does not require product specific calibration.

Ramanand Singh Yadav, Ralf Metzler, Rajarshi Chakrabarti

Using computer simulations in two dimensions we investigate the dynamics and structure of passive polymeric tracer with different topologies immersed in a low-density active particle bath. One of the key observations is that polymer exhibit faster dynamics compared to passive colloidal particles at high activity, for the same particle density, in both linear

Anthony Genevois, Anne Lonjou, Christian Urech

This article is dedicated to the computation of an explicit presentation of some asymptotically rigid mapping class groups, namely the braided Higman-Thompson groups. To do so, we use the action of these groups on the spine complex, a simply connected cube complex constructed by the authors in a previous work. In particular, this allows to compute the abelia

Mayank Nagda, Phil Ostheimer, Justus Arweiler, Indra Jungjohann

Style transfer combines the content of one signal with the style of another. It supports applications such as data augmentation and scenario simulation, helping machine learning models generalize in data-scarce domains. While well developed in vision and language, style transfer methods for time series data remain limited. We introduce DiffTSST, a diffusion-

Fabio Ancona, Mohamed Bentaibi, Francesco Rossi

Finding conditions ensuring consensus, i.e. convergence to a common value, for a networked system is of crucial interest, both for theoretical reasons and applications. This goal is harder to achieve when connections between agents are temporarily lost. Here, we prove that known conditions (introduced by Moreau) ensure an exponential convergence to consensus

A. A. Kamashev, N. N. Garif'yanov, A. A. Validov, A. S. Osin

The structures of the superconducting spin valve (SSV) Fe/Si$_3$N$_4$/Pb/Si$_3$N$_4$/Fe (where Si$_3$N$_4$ is a dielectric insulating layer of controlled thickness) were investigated. The dependence of the magnitude of the SSV effect on the thicknesses of the superconducting (S) and insulating (I) layers was studied. Optimization of the S and I layer thickne

Geivison Ribeiro

We revisit the results of Kitson and Timoney \emph{[J.~Math.~Anal.~Appl.~\textbf{378} (2011), 680--686]} on the spaceability of complements of operator ranges, extending one of their main theorems to the general Fr\'echet setting. In particular, we provide an affirmative answer to the question posed in \emph{Remark~3.4} of that paper, showing that the conclu

Bingjie Zhu, Zhixiong Chen, Liqiang Zhao, Hyundong Shin

Large language model (LLM) inference at the network edge is a promising serving paradigm that leverages distributed edge resources to run inference near users and enhance privacy. Existing edge-based LLM inference systems typically adopt autoregressive decoding (AD), which only generates one token per forward pass. This iterative process, compounded by the l

KiHyun Nam, Jongmin Choi, Hyeongkeun Lee, Jungwoo Heo

Contrastive audio-language pretraining yields powerful joint representations, yet a persistent audio-text modality gap limits the benefits of coupling multimodal encoders with large language models (LLMs). We present Diffusion-Link, a diffusion-based modality-bridging module that generatively maps audio embeddings into the text-embedding distribution. The mo

Ming-Ming Du Yi-Hao Fan, Hong-Wei Li, Shu-Ting Shen, Xiao-Jing Yan

We investigate the behavior of basis-independent quantum coherence between two modes of a free Dirac field as observed by relatively accelerated observers. Our findings reveal three key results: (i) the basis-independent coherence between modes A and BI decreases with increasing acceleration but remains finite even in the limit of infinite acceleration; (ii)

Chenxi Wang, Yixuan Zhang, Ruiji Yu, Yufei Zheng

As the demand for emotional intelligence in large language models (LLMs) grows, a key challenge lies in understanding the internal mechanisms that give rise to emotional expression and in controlling emotions in generated text. This study addresses three core questions: (1) Do LLMs contain context-agnostic mechanisms shaping emotional expression? (2) What fo

Dennis Müller, Maurice Chiodo, Michael Meyer

The literature on ethics and sustainability in mathematics and its education is increasingly complex and fragmented, potentially leading to communication breakdowns between different scholarly traditions. To address this, the paper introduces the "Ethical and Sustainable Concerns Triangle," a framework that maps discourses based on their relative concern for

Namrata Roy, Alaina Henry, Tucker Jones, Ivana Barisic

We present spatially resolved rest-optical spectroscopy of 38 star-forming galaxies at 0.5 < z < 1.7 from the JWST/NIRSpec MSA-3D survey, which uses slit-stepping to build IFU-like datacubes at 0.1'' resolution. We map emission-line morphology, excitation, and kinematics of the warm ionized gas using [N II]/H$\alpha$, [S II]/H$\alpha$, and [O III]/H$\beta$.

J. M. Marcos

Out-of-equilibrium systems, inherently complex and challenging to understand, are prevalent across various disciplines, including physics where they arise in contexts such as fluid dynamics. In particular, critical out-of-equilibrium systems combine this complexity with the scaling laws and universality classes observed in critical phenomena, with kinetic su

Sarah Rebecca Ondraszek, Jörg Waitelonis, Katja Keller, Claudia Niessner

An essential component for evaluating and comparing physical and cognitive capabilities between populations is the testing of various factors related to human performance. As a core part of sports science research, testing motor performance enables the analysis of the physical health of different demographic groups and makes them comparable. The Motor Resear

Zhe Wang, Yaming Yang, Ziyu Guan, Bin Tong

In recent years, affiliate marketing has emerged as a revenue-sharing strategy where merchants collaborate with promoters to promote their products. It not only increases product exposure but also allows promoters to earn a commission. This paper addresses the pivotal yet under-explored challenge in affiliate marketing: accurately assessing and predicting th

Alice L. L. Gao, Yun Li, Matthew H. Y. Xie

In this paper, we focus on the equivariant inverse Kazhdan--Lusztig polynomials of thagomizer matroids, a natural family of graphic matroids associated with the complete tripartite graphs $K_{1,1,n}$. These polynomials were introduced by Proudfoot as an extension of the Kazhdan--Lusztig theory for matroids. We derive closed-form expressions for the $\mathfra

Ruizhe Liu, Pei Zhou, Qian Luo, Li Sun

Effective generalization in robotic manipulation requires representations that capture invariant patterns of interaction across environments and tasks. We present a self-supervised framework for learning hierarchical manipulation concepts that encode these invariant patterns through cross-modal sensory correlations and multi-level temporal abstractions witho

Kazuki Watanabe, Mayuko Kori, Taro Sekiyama, Satoshi Kura

We propose a categorical framework for linear-time temporal verification of effectful higher-order programs, including probabilistic higher-order programs. Our framework provides a generic denotational reduction -- namely, a denotational product construction -- from linear-time safety verification of effectful higher-order programs to computation of weakest

Haximjan Abdusattar

We investigate $P$-$V$ criticality, with a focus on Ruppeiner geometry, in the extended phase space of Hayward anti-de Sitter (AdS) black holes. Through thermodynamic analysis, we confirm that Hayward-AdS black holes undergo distinct $P$-$V$ phase transitions and exhibit well-defined critical phenomena in the vicinity of their critical points. These behavior

Lubomíra Dvořáková, Edita Pelantová, Jeffrey Shallit

In 2017, Clark Kimberling defined an interesting sequence ${\bf B} = 0100101100 \cdots$ of $0$'s and $1$'s by certain inflation rules, and he made a number of conjectures about this sequence and some related ones. In this note we prove his conjectures using, in part, the Walnut theorem-prover. We show how his word is related to the infinite Tribonacci word,

Munsif Ali, Leonardo Rossi, Massimo Bertozzi

Continual learning (CL) in the context of Generative Adversarial Networks (GANs) remains a challenging problem, particularly when it comes to learn from a few-shot (FS) samples without catastrophic forgetting. Current most effective state-of-the-art (SOTA) methods, like LFS-GAN, introduce a non-negligible quantity of new weights at each training iteration, w

Chen Gao, Zixin Zhao, Lv Shao, Tong Liu

Click-Through Rate (CTR) prediction has long been dominated by discriminative paradigms that optimize local decision boundaries within candidate-specific subspaces. However, these models often fail to capture the global joint distribution and the continuous structural evolution of user intent across all-domain movelines. While generative approaches attempt t

Kaj Munhoz Arfvidsson, Loizos Hadjiloizou, Frank J. Jiang, Karl H. Johansson

In this paper, we present pyspect, a Python toolbox that simplifies the use of reachability analysis for temporal logic problems. Currently, satisfying complex requirements in cyber-physical systems requires significant manual effort and domain expertise to develop the underlying reachability programs. This high development effort limits the broader adoption

Pankaj Kumar, Vivek Vijay

Heavy-tailed probability distributions are extremely useful and play a crucial role in modeling different types of financial data sets. This study presents a two-pronged methodology. First, a mixture probability distribution is created by combining Gaussian and Rayleigh distributions using the arctangent transformation, aimed at producing heavier-tailed feat

Belkiss Souayed, Sarah Ebling, Yingqiang Gao

Individuals with intellectual disabilities often have difficulties in comprehending complex texts. While many text-to-image models prioritize aesthetics over accessibility, it is not clear how visual illustrations relate to text simplifications (TS) generated from them. This paper presents a structured vision-language model (VLM) prompting framework for gene

Le Ngoc Luyen, Marie-Hélène Abel

This paper investigates automated skill decomposition using Large Language Models (LLMs) and proposes a rigorous, ontology-grounded evaluation framework. Our framework standardizes the pipeline from prompting and generation to normalization and alignment with ontology nodes. To evaluate outputs, we introduce two metrics: a semantic F1-score that uses optimal

Konstantinos Oikonomidis, Jan Quan, Panagiotis Patrinos

We study nonlinearly preconditioned gradient methods for smooth nonconvex optimization problems, focusing on sigmoid preconditioners that inherently perform a form of gradient clipping akin to the widely used gradient clipping technique. Building upon this idea, we introduce a novel heavy ball-type algorithm and provide convergence guarantees under a general

Micha Christoph, Barnabás Janzer, Kalina Petrova, Raphael Steiner

A famous conjecture by Thomassen from 1983 asserts that for any given $k,g\in \mathbb{N}$ there exists some $d=d(k,g)\in \mathbb{N}$ such that every graph of minimum degree at least $d$ contains a subgraph of minimum degree at least $k$ and girth at least $g$. In this paper, we initiate the systematic study of the directed analogs of Thomassen's conjecture o

Shinhyung Yang, David Georg Reichelt, Henrik Ingo, Wilhelm Hasselbring

In GitHub with its 518 million hosted projects, performance changes within these projects are highly relevant to the project's users. Although performance measurement is supported by GitHub CI/CD, performance change detection is a challenging topic. In this paper, we demonstrate how we incorporated Nyrki\"o to MooBench. Prior to this work, Moobench continuou

Angela Maria Fajardo Lacave, Francesco Vicentini, Fabian Welschinger, Laura De Lorenzis

We propose a novel variational phase-field model for fracture in anisotropic materials. The model is specifically designed to allow a more flexible calibration of crack nucleation than existing anisotropic fracture formulations, while avoiding the introduction of multiple damage variables. In addition to the classical components of anisotropic phase-field mo

Weixi Situ, Hanjing Ye, Jianwei Peng, Yu Zhan

Robot person following (RPF) is a core capability in human-robot interaction, enabling robots to assist users in daily activities, collaborative work, and other service scenarios. However, achieving practical RPF remains challenging due to frequent occlusions, particularly in dynamic and crowded environments. Existing approaches often rely on fixed-point fol

Sabrina McCallum, Amit Parekh, Alessandro Suglia

Current approaches to embodied AI tend to learn policies from expert demonstrations. However, without a mechanism to evaluate the quality of demonstrated actions, they are limited to learning from optimal behaviour, or they risk replicating errors and inefficiencies. While reinforcement learning offers one alternative, the associated exploration typically re

Xiaobin Zhou, Miao Wang, Chengao Li, Can Cui

Rotor failures in quadrotors may result in high-speed rotation and vibration due to rotor imbalance, which introduces significant challenges for autonomous flight in unknown environments. The mainstream approaches against rotor failures rely on fault-tolerant control (FTC) and predefined trajectory tracking. To the best of our knowledge, online failure detec

Jean-Paul Travert, Cédric Goeury, Sébastien Boyaval, Vito Bacchi

Flood mapping and water depth estimation from Synthetic Aperture Radar (SAR) imagery are crucial for calibrating and validating hydraulic models. This study uses SAR imagery to evaluate various preprocessing (especially speckle noise reduction), flood mapping, and water depth estimation methods. The impact of the choice of method at different steps and its h

Daniel Radmanovac, Andreas Gsponer, Simon Waid, Sebastian Onder

Silicon carbide (SiC) planar PiN diodes from two different manufacturers were irradiated with 252.7 MeV protons from a medical synchrotron. Over the course of two 8h irradiation shifts, the samples were exposed to increasing fluences ranging from 1.4e+11 to 3.5e+13 p+/cm^2. Electrical characterizations, including IV and CV measurements, were performed both b

Samer Al-Hamadani

Object detection traditionally relies on costly manual annotation. We present the first comprehensive cost-effectiveness analysis comparing supervised YOLO and zero-shot vision-language models (Gemini Flash 2.5 and GPT-4). Evaluated on 5,000 stratified COCO images and 500 diverse product images, combined with Total Cost of Ownership modeling, we derive break

Junhua Zhou, Quanjun Li, Weixuan Li, Guang Yu

The rise of digital medical imaging, like MRI and CT, demands strong encryption to protect patient data in telemedicine and cloud storage. Chaotic systems are popular for image encryption due to their sensitivity and unique characteristics, but existing methods often lack sufficient security. This paper presents the Three-dimensional Diffusion Algorithm and

Bernd Hofmann, Niklas Piechulek, Sven Kreitlein, Joerg Franke

This paper proposes an agent-based approach toward a more natural interface between humans and machines. Large language models equipped with tools and the communication standard OPC UA are utilized to control machines in natural language. Instead of touch interaction, which is currently the state-of-the-art medium for interaction in operations, the proposed

Josep Domingo-Ferrer, David Sánchez

Privacy models were introduced in privacy-preserving data publishing and statistical disclosure control with the promise to end the need for costly empirical assessment of disclosure risk. We examine how well this promise is kept by the main privacy models. We find they may fail to provide adequate protection guarantees because of problems in their definitio

Andreas Krug, Fabian Reede, Ziyu Zhang

We construct new stable vector bundles on Hilbert schemes of points on algebraic surfaces, which are parametrised by connected components of their moduli spaces. This work generalises aspects of our previous work on tautological bundles and of recent work of O'Grady.

Ruirui Chen, Weifeng Jiang, Chengwei Qin, Bo Xiong

Knowledge graphs (KGs) are widely used in knowledge-intensive applications, yet it remains unclear how effectively current large language models (LLMs) can construct document-grounded KGs in a zero-shot, schema-free setting without relying on complex task-specific frameworks. We introduce Detail-to-Abstract Hierarchical Knowledge Graph (D2A-HKG) construction

Jian Lan, Zhicheng Liu, Udo Schlegel, Raoyuan Zhao

Large vision-language models (VLMs) achieve strong performance in Visual Question Answering but still rely heavily on supervised fine-tuning (SFT) with massive labeled datasets, which is costly due to human annotations. Crucially, real-world datasets often exhibit human uncertainty (HU) -- variation in human confidence across annotations -- but standard SFT

Run Gu, Renjie Xie, Wei Xu, Zhaohui Yang

Superimposed pilot (SIP) schemes face significant challenges in effectively superimposing and separating pilot and data signals, especially in multiuser mobility scenarios with rapidly varying channels. To address these challenges, we propose a novel channel-aware learning framework for SIP schemes, termed CaSIP, that jointly optimizes pilot-data power (PDP)

Bahareh Afshari, Johannes Kloibhofer

We present a syntactic cut-elimination procedure for the alternation-free fragment of the modal mu-calculus. Cut reduction is carried out within a cyclic proof system, where proofs are finitely branching but may be non-wellfounded. The structure of such proofs is exploited to directly transform a cyclic proof with cuts into a cut-free one, without detouring

Wenbo Wu, Qingyi Si, Xiurui Pan, Ye Wang

While Key-Value (KV) cache succeeds in reducing redundant computations in auto-regressive models, it introduces significant memory overhead, limiting its practical deployment in long-sequence scenarios. Existing KV retrieval methods mitigate this by dynamically retaining only a subset of KV entries on the GPU. However, they still suffer from notable efficien

Muhammad Bilal, Xiaolong Xu

This work presents a Network-Optimised Spiking (NOS) delay-aware scheduler for 6G radio access. The scheme couples a bounded two-state kernel to a clique-feasible proportional-fair (PF) grant head: the excitability state acts as a finite-buffer proxy, the recovery state suppresses repeated grants, and neighbour pressure is injected along the interference gra

Miloš Ciganović, Elena Scola Gagliardi, Massimiliano Tancioni

We estimate the dynamic distributional effects of financial shocks in the Euro Area using survey-based microdata on personal incomes. We find that positive financial shocks increase inequality, with heterogeneity across different income groups. Much of the response emerges in the tails of the income distribution. By decomposing individual incomes into financ

Nikita Afonin, Nikita Andriianov, Vahagn Hovhannisyan, Nikhil Bageshpura

Recent work has shown that narrow finetuning can produce broadly misaligned LLMs, a phenomenon termed emergent misalignment (EM). While concerning, these findings were limited to finetuning and activation steering, leaving out in-context learning (ICL). We therefore ask: does EM emerge in ICL? We find that it does: across four model families (Gemini, Kimi-K2

Han Xia, Quanjun Li, Qian Li, Zimeng Li

Medical image segmentation is vital for diagnosis, treatment planning, and disease monitoring but is challenged by complex factors like ambiguous edges and background noise. We introduce EEMS, a new model for segmentation, combining an Edge-Aware Enhancement Unit (EAEU) and a Multi-scale Prompt Generation Unit (MSPGU). EAEU enhances edge perception via multi

Jack Jackman, David Ryan, Arun Narayanan, Pedro Nardelli

Modern power grids face an acute mismatch between where data is generated and where it can be processed: protection relays, EV (Electric Vehicle) charging, and distributed renewables demand millisecond analytics at the edge, while energy-hungry workloads often sit in distant clouds leading to missed real-time deadlines and wasted power. We address this by pr

Pablo Tieben, Jan Rhensius, Takuya F. Segawa, Risei Abe

Diamonds containing color centers have recently gathered significant attention for photonic quantum technologies, including quantum sensing, photonic quantum computers, and quantum networks. Among the various color centers, tin-vacancy (SnV) centers are particularly promising due to the high emission efficiency from the zero-phonon line and due to their long

Virgile Troude, Didier Sornette

Heavy-tailed fluctuations and power law statistics pervade physics, finance, and economics, yet their origin is often ascribed to systems poised near criticality. Here we show that such behavior can emerge far from instability through a universal mechanism of non-normal eigenvector amplification in multidimensional Kesten processes $x_{t+1}=A_t x_t+\eta_t$,

Vassilios Karakostas, Elias Zafiris

Contemporary perspectivism is viewed as a framework of scientific inquiry concerning the origin, generation and systematization of scientific knowledge of nature by focusing on the conditions under which such knowledge may arise in perspectivist terms and investigating the essential ramifications of these conditions. To this end, we develop the conceptual, m