November 2025 arXiv papers — page 109

Mattias Brännström, Themis Dimitra Xanthopoulou, Lili Jiang

Avoiding bias and understanding the real-world consequences of AI-supported decision-making are critical to address fairness and assign accountability. Existing approaches often focus either on technical aspects, such as datasets and models, or on high-level socio-ethical considerations - rarely capturing how these elements interact in practice. In this pape

Oleg Davydov

We provide improved error bounds for kernel-based numerical differentiation in terms of growth functions when kernels are of a finite smoothness, such as polyharmonic splines, thin plate splines or Wendland kernels. In contrast to existing literature, the new estimates take into account the H\"older class smoothness of kernel's derivatives, which helps to im

Masoud Ganji, Cristina Giannotti, Andrea Spiro

Building on the characterisation in [C. D. Hill, J. Lewandowski and P. Nurowski, Indiana Univ. Math. J. 57 (2008), 3131--3176] of 4-dimensional Lorentzian metrics adapted to an optical structure and satisfying the null fluid Einstein equations, we give an explicit parameterisation of this class under the assumption that the optical structure is of Kerr type.

Karim Khamaisi, Pascal Kiechl, Katharina Müller, Burkhard Stiller

Pulse-wave Distributed Denial-of-Service (DDoS) attacks generate short, synchronized bursts of traffic that circumvent pattern-based detection and quickly exhaust traditional defense systems. This transient and spatially distributed behavior makes analysis extremely challenging, as no public datasets capture how such attacks evolve across multiple network do

Andreas Thom

In the analysis of three-dimensional biological microstructures such as organoids, microscopy frequently yields two-dimensional optical sections without access to their orientation. Motivated by the question of whether such random planar sections determine the underlying three-dimensional structure, we investigate a discrete analogue in which the ambient str

Stephan Nef, Bruno Rodrigues

Digital mental-health sensing increasingly depends on mobile or wearable devices that require intrusive permissions and continuous user compliance. We present CareNet, a router-centric system that transforms household network metadata into interpretable behavioral indicators aligned with DSM-5 depressive-symptom domains. All processing occurs locally at the

Anna Wienhard

In this article we revisit a new notion of positivity in real semisimple Lie groups that at the same time generalizes total positivity in split real Lie groups as well as positive Lie semigroups in Hermitian Lie groups of tube type. We shortly discuss the relationship with higher rank Teichm\"uller spaces, and then focus on describing different aspects of po

Zhenyu Lei, Patrick Soga, Yaochen Zhu, Yinhan He

Understanding and continuously refining multimodal molecular knowledge is crucial for advancing biomedicine, chemistry, and materials science. Molecule language models (MoLMs) have become powerful tools in these domains, integrating structural representations (e.g., SMILES strings, molecular graphs) with rich contextual descriptions (e.g., physicochemical pr

Luyao Niu, Zepu Wang, Shuyi Guan, Yang Liu

While spatio-temporal Graph Neural Networks (GNNs) excel at modeling recurring traffic patterns, their reliability plummets during non-recurring events like accidents. This failure occurs because GNNs are fundamentally correlational models, learning historical patterns that are invalidated by the new causal factors introduced during disruptions. To address t

Noah Hong, Tao Hong

Phase transitions have been proposed as the origin of emergent abilities in large language models (LLMs), where new capabilities appear abruptly once models surpass critical thresholds of scale. Prior work, such as that of Wei et al., demonstrated these phenomena under model and data scaling, with transitions revealed after applying a log scale to training c

Morrison Turnansky

We present the Lukasiewicz logic as a viable system for an implication algebra on a system of qubits. Our results show that the three valued Lukasiewicz logic can be embedded in the stabilized space of an arbitrary quantum error correcting stabilizer code. We then fully characterize the non trivial errors that may occur up to group isomorphism. Lastly, we de

Cătălin-Alexandru Rîpanu, Andrei-Theodor Hotnog, Giulia-Stefania Imbrea, Dumitru-Clementin Cercel

Automatic sign language recognition plays a crucial role in bridging the communication gap between deaf communities and hearing individuals; however, most available datasets focus on American Sign Language. For Romanian Isolated Sign Language Recognition (RoISLR), no large-scale, standardized dataset exists, which limits research progress. In this work, we i

Chaouki Hjaiji, Bassem Ouni, Mohamed-Slim Alouini

High-Altitude Platform Stations (HAPS) are emerging stratospheric nodes within non-terrestrial networks. We provide a structured overview of HAPS subsystems and principal communication links, map cybersecurity and privacy exposure across communication, control, and power subsystems, and propose a stratosphere-aware threat taxonomy. We then discuss defenses f

Herbert Baaser

In this paper, we describe a uniform and standardized approach for analytically verifying the stability of isotropic, incompressible hyperelastic material models. Here, we address {\sl stability} as fulfillment of the {\sc Hill} condition -- i.e.\ the positive definiteness of the material modulus in the {\sc Kirchhoff} stress -- log--strain relation. For inc

Stephanie Atherton, Somak Dutta, Jordy Lopez Garcia, Joel Louwsma

We describe a significant update to the Macaulay2 package A1BrouwerDegrees. We extend several methods in the previous version of the package to the setting of finite \'{e}tale algebras, allowing the computation of transfers along finite \'{e}tale extensions. Additionally, we implement a number of new features for the computation of unstable $\mathbb{A}^{1}$-

Hao Wei, Aleksandra Franz, Bjoern List, Nils Thuerey

When simulating partial differential equations, hybrid solvers combine coarse numerical solvers with learned correctors. They promise accelerated simulations while adhering to physical constraints. However, as shown in our theoretical framework, directly applying learned corrections to solver outputs leads to significant autoregressive errors, which originat

Harrison Katz

This commentary translates the central ideas in Lead times in flux into a practice ready handbook in R. The original article measures change in the full distribution of booking lead times with a normalized L1 distance and tracks that divergence across months relative to year over year and to a fixed 2018 reference. It also provides a bound that links diverge

John R. Thorstensen, Annabelle E. Niblett, Shreya Gandhi, Lauren P. Zanarini

We present observations and analyses of five little-studied cataclysmic binary stars in the southern celestial hemisphere. Our new observations are from the South African Astronomical Observatory. The objects and salient results are as follows: (i) 6dF0752-54 is a dwarf nova with an orbital period Porb = 5.05 hr that shows a contribution from a mid-M type se

Hanna Furmańczyk, Didem Gözüpek, Sibel Özkan

The packing coloring problem has diverse applications, including frequency assignment in wireless networks, resource distribution and facility location in smart cities and post-disaster management, as well as in biological diversity. Formally, the packing coloring of a graph is a vertex coloring in which any two vertices assigned color $i$ are at a distance

Ben Gao, Jordan Patracone, Stéphane Chrétien, Olivier Alata

We introduce Conformal Online Learning of Koopman embeddings (COLoKe), a novel framework for adaptively updating Koopman-invariant representations of nonlinear dynamical systems from streaming data. Our modeling approach combines deep feature learning with multistep prediction consistency in the lifted space, where the dynamics evolve linearly. To prevent ov

James Moore

Human memory retrieval often resembles ecological foraging where animals search for food in a patchy environment. Optimal foraging means following the Marginal Value Theorem (MVT), in which individuals exploit a patch of semantically related concepts until it becomes less rewarding and then switch to a new cluster. While human behavioral data suggests foragi

Shih-Chi Liao, Maziar S. Hemati, Peter Seiler

Boundedness is an important property of many physical systems. This includes incompressible fluid flows, which are often modeled by quadratic dynamics with an energy-preserving nonlinearity. For such systems, Schlegel and Noack proposed a sufficient condition for boundedness utilizing quadratic Lyapunov functions. They also propose a necessary condition for

Nicholas Karris, Luke Durell, Javier Flores, Tegan Emerson

It can be shown that Stable Diffusion has a permutation-invariance property with respect to the rows of Contrastive Language-Image Pretraining (CLIP) embedding matrices. This inspired the novel observation that these embeddings can naturally be interpreted as point clouds in a Wasserstein space rather than as matrices in a Euclidean space. This perspective o

Sungjun Seo, Kooktae Lee

This paper addresses the fundamental problem of non-uniform area coverage in multi-agent systems, where different regions require varying levels of attention due to mission-dependent priorities. Existing uniform coverage strategies are insufficient for realistic applications, and many non-uniform approaches either lack optimality guarantees or fail to incorp

Sören Dréano, Derek Molloy, Noel Murphy

This work introduces Llamazip, a novel lossless text compression algorithm based on the predictive capabilities of the LLaMA3 language model. Llamazip achieves significant data reduction by only storing tokens that the model fails to predict, optimizing storage efficiency without compromising data integrity. Key factors affecting its performance, including q

Aleesha Khurram, Amir Moeini, Shangtong Zhang, Rohan Chandra

Despite significant progress and advances in autonomous driving, many end-to-end systems still struggle with domain adaptation (DA), such as transferring a policy trained under clear weather to adverse weather conditions. Typical DA strategies in the literature include collecting additional data in the target domain or re-training the model, or both. Both th

Benjamin Li, Shuyang Shi, Lucia Romero, Huao Li

Adaptation is the cornerstone of effective collaboration among heterogeneous team members. In human-agent teams, artificial agents need to adapt to their human partners in real time, as individuals often have unique preferences and policies that may change dynamically throughout interactions. This becomes particularly challenging in tasks with time pressure

Léopold Van Brandt, Grégoire Brandsteert, Denis Flandre

The excitability property of spiking neurons describes their capability to output an action potential as a real-time response to an input synaptic excitation current and is central to the event-based neuromorphic computing paradigm. The spiking mechanism is analysed in a representative ultra-low-power analog neuron from the circuit literature. Relying on con

Mukkesh Ganesh, Kaushik Iyer, Arun Baalaaji Sankar Ananthan

The Key Value(KV) cache is an important component for efficient inference in autoregressive Large Language Models (LLMs), but its role as a representation of the model's internal state makes it a potential target for integrity attacks. This paper introduces "History Swapping," a novel block-level attack that manipulates the KV cache to steer model generation

Timur Anvar, Jeffrey Chen, Yuyan Wang, Rohan Chandra

Autonomous vehicle navigation in complex environments such as dense and fast-moving highways and merging scenarios remains an active area of research. A key limitation of RL is its reliance on well-specified reward functions, which often fail to capture the full semantic and social complexity of diverse, out-of-distribution situations. As a result, a rapidly

Ahmed Hussain, Asmaa Abdallah, Abdulkadir Celik, Ahmed M. Eltawil

Recent studies suggest that uniform circular arrays (UCAs) can extend the angular coverage of the radiative near field region. This work investigates whether such enhanced angular coverage translates into improved spatial multiplexing performance when compared to uniform linear arrays (ULAs). To more accurately delineate the effective near field region, we i

Zong-Han Bai, Po-Yen Chu

Intermittent demand forecasting poses unique challenges due to sparse observations, cold-start items, and obsolescence. Classical models such as Croston, SBA, and the Teunter--Syntetos--Babai (TSB) method provide simple heuristics but lack a principled generative foundation. We introduce TSB-HB, a hierarchical Bayesian extension of TSB. Demand occurrence is

Phan Thành Nam, Simone Rademacher, Avy Soffer

We consider the Bogoliubov approximation for the many-body quantum dynamics of weakly interacting Bose gases and establish a uniform-in-time validity of the Bogoliubov theory. The proof relies on a detailed analysis of the dispersive behavior of the symplectic Bogoliubov dynamics, which allows for a rigorous derivation of the Bogoliubov theory as an effectiv

Aritro Pathak

For a linear elliptic operator with a singular drift that satisfies a finite Carleson measure condition, we prove that there exist `ample' sawtooth domains of the unit ball $B(0,1)\subset \R^{n+1}$ so that a BMO solvability assumption in these sawtooth subdomains implies that the elliptic measure satisfies the weak $A_\infty$ condition with respect to th

Filippo Caleca, Saverio Bocini, Fabio Mezzacapo, Tommaso Roscilde

Finite-size quantum spin systems can be magnetized by the application of a symmetry-breaking field, but in general their symmetry is expected to be restored once the field is turned off adiabatically. Recently (F. Caleca et al., arXiv:2412.15493) we have shown that systems of half-integer spins with an odd number of sites and a parity-preserving Hamiltonian

Vivek Chawla, Boris Slautin, Utkarsh Pratiush, Dayakar Penumadu

Scientific datasets often arise from multiple independent mechanisms such as spatial, categorical or structural effects, whose combined influence obscures their individual contributions. We introduce DIVIDE, a framework that disentangles these influences by integrating mechanism-specific deep encoders with a structured Gaussian Process in a joint latent spac

Colton R. Crum, Christopher Sweet, Adam Czajka

Integrating human perceptual priors into the training of neural networks has been shown to raise model generalization, serve as an effective regularizer, and align models with human expertise for applications in high-risk domains. Existing approaches to integrate saliency into model training often rely on internal model mechanisms, which recent research sugg

Adrita Rahman Tori, Khondokar Fida Hasan

The performance of Machine Learning (ML) and Deep Learning (DL)-based Intrusion Detection and Prevention Systems (IDS/IPS) is critically dependent on the relevance and quality of the datasets used for training and evaluation. However, current AI model evaluation practices for developing IDS/IPS focus predominantly on accuracy metrics, often overlooking wheth

Zhongteng Cai, Yaxuan Wang, Yang Liu, Xueru Zhang

As synthetic data proliferates across the Internet, it is often reused to train successive generations of generative models. This creates a ``self-consuming loop" that can lead to training instability or \textit{model collapse}. Common strategies to address the issue -- such as accumulating historical training data or injecting fresh real data -- either incr

Aritro Pathak

We introduce a technique to obtain pointwise upper and lower bounds for the Green's function of elliptic operators whose principal part is the Laplacian and that include a drift term diverging near the boundary like a power of the inverse distance with exponent less than 1, in the unit ball B(0,1) \subset \mathbb{R}^n, n \ge 3. The constants in the upper est

Amirhossein Hassanzadeh, Bartosz Krawczyk, Michael Saunders, Rob Wible

Voxelization is an effective approach to reduce the computational cost of processing Light Detection and Ranging (LiDAR) data, yet it results in a loss of fine-scale structural information. This study explores whether low-level voxel content information, specifically target occupancy percentage within a voxel, can be inferred from high-level voxelized LiDAR

Yaniv Hacmon, Keren Gorelik, Gilad Gressel, Yisroel Mirsky

Fingerprint recognition systems are widely deployed for authentication and forensic applications, but the security of stored fingerprint data remains a critical vulnerability. While many systems avoid storing raw fingerprint images in favor of minutiae-based templates, recent research shows that these templates can be reverse-engineered to reconstruct realis

Jonatas Grosman, Cassio Almeida, Guilherme Schardong, Hélio Lopes

Using representations provided by a large pre-trained model has become the primary strategy for achieving state-of-the-art results in a wide range of tasks. A recently proposed large pre-trained model, wav2vec 2.0, was seminal for several other works on pre-training large models on speech data. Many models are being pre-trained using the same architecture as

Adiba Amira Siddiqa, Sayed Shafaat Mahmud, Rafael Martinez-Galarza

We introduce a Variational Autoencoder (VAE)--Normalizing Flow (NF) framework for rapid probabilistic inference of galaxy properties and emission line fluxes at $z \leq 0.3$ from SDSS \textit{gri} imaging and photometry. Our model probabilistically infers stellar mass, star formation rate (SFR), redshift, gas-phase metallicity, and central black hole mass fo

Osvaldo Guzmán, Carlos López-Callejas

We generalize the diamond principle and its variants using the notion of stationarity in trees introduced by Brodsky in [Brodsky, A. M., A theory of stationary trees and the balanced Baumgartner--Hajnal--Todorcevic theorem for trees. The Bulletin of Symbolic Logic]. In particular, we show that if $T$ is a nonspecial $\omega_1$-tree, then $\diamondsuit_T \imp

Ankita Raj, Chetan Arora

Open-vocabulary object detectors (OVODs) unify vision and language to detect arbitrary object categories based on text prompts, enabling strong zero-shot generalization to novel concepts. As these models gain traction in high-stakes applications such as robotics, autonomous driving, and surveillance, understanding their security risks becomes crucial. In thi

Adel Hidri, Suleiman Ali Alsaif, Muteeb Alahmari, Eman AlShehri

Understanding customer attitudes has become a critical component of decision-making due to the growing influence of social media and e-commerce. Text-based opinions are the most structured, hence playing an important role in sentiment analysis. Most of the existing methods, which include lexicon-based approaches and traditional machine learning techniques, a

Sadruddin Rahimi, Saeid Alikhani

This paper explores the Harmonic matrix $MH(G)$ associated with a simple graph $ G $, where each entry corresponds to $ \frac{2}{d_i + d_j} $ for adjacent vertices $ v_i $ and $ v_j $. We investigate the spectral properties of this matrix, particularly focusing on its eigenvalues. A central objective of this work is to compute the Harmonic characteristic pol

Ahmed Hussain, Ahmed Sultan, Asmaa Abdallah, Abdulkadir Celik

Near-field beamforming enables target discrimination in both range (axial) and angle (lateral) dimensions. Elevated sidelobes along either dimension, however, increase susceptibility to interference and degrade detection performance. Conventional amplitude tapering techniques, designed for far-field scenarios, cannot simultaneously suppress axial and lateral

Domenico Di Sante, Titus Neupert, Giorgio Sangiovanni, Ronny Thomale

Three important driving forces for creating qualitatively new phases in quantum materials are the topology of the materials' electronic band structures, frustration in the electrons' motion or magnetic interactions, and strong correlations between their charge, spin, and orbital degrees of freedom. In very few material systems do all of these aspects come to

Emilien Valat, Ozan Öktem

Ensuring proper generalization is a critical challenge in applying data-driven methods for solving inverse problems in imaging, as neural networks reconstructing an image must perform well across varied datasets and acquisition geometries. In X-ray Computed Tomography (CT), convolutional neural networks (CNNs) are widely used to filter the projection data bu

Amirhossein Hassanzadeh, Robert Mancini, Aaron Gerace, Rehman Eon

Current Landsat Level 2 surface temperature products are derived using a single-channel (SC) methodology to estimate per-pixel surface temperature (ST) maps from Level~1 radiance data. A known issue with the Level 2 uncertainty, however, is its susceptibility to overestimation of uncertainty due to its dependence on Landsat's cloud mask, which is prone to fa

Lea Hergert, Gábor Berend, Mario Szegedy, Gyorgy Turan

Large language models (LLMs) achieve superhuman performance on complex reasoning tasks, yet often fail on much simpler problems, raising concerns about their reliability and interpretability. We investigate this paradox through a focused study with two key design features: simplicity, to expose basic failure modes, and scale, to enable comprehensive controll

Patrick Barlatier, Richard Dapoigny

Qualitative spatial models based on Goodman-style mereology and pseudo-topology often pose problems for advanced geometric reasoning, as they lack true Euclidean geometry and fully developed topological spaces. We address this issue by extending an existing formalization grounded in a dependent type theory using the Coq proof assistant, together with a White

Philip Soliman, Filipe Cumaru, Alexander Heinlein

A new iteration bound for the preconditioned conjugate gradient (PCG) method is presented that more accurately captures convergence for systems with clustered eigenspectra, where the classical condition number-based bound is too pessimistic. By using the edge eigenvalues of each cluster in the spectral distribution, the bound is shown to be orders of magnitu

William Ward Armstrong, Hongyi Li, Jun Xu

A method for creating a forest of model trees to fit samples of a function defined on images is described in several steps: down-sampling the images, determining a tree's hyperplanes, applying convolutions to the hyperplanes to handle small distortions of training images, and creating forests of model trees to increase accuracy and achieve a smooth fit. A 1-

Nazmiye Alemdar, Hürmet Fulya Akız, Halim Ayaz

In this study, the soft usual topology compatible with the usual topology of $\mathbb{R}$ is defined, and using its subspace topology on the interval $[0,1]$, the concept of a soft path is introduced. Within this context, the notions of soft connectedness and soft path connectedness are developed, their relationship is analyzed, and it is shown that these pr

Nakshatra Gupta, Sumanth Prabhu, Supratik Chakraborty, R Venkatesh

Data poisoning is a training-time attack that undermines the trustworthiness of learned models. In a targeted data poisoning attack, an adversary manipulates the training dataset to alter the classification of a targeted test point. Given the typically large size of training dataset, manual detection of poisoning is difficult. An alternative is to automatica

Miguel Castillo-Celeita, Matteo Schiavon

This work presents a study of continuous-variable quantum key distribution (CV-QKD) protocols over fast-fading channels, typically found in free-space communication links. Two eavesdropping models are considered to evaluate their security under collective attacks: \textit{Holevo bound average} (HBA) and \textit{covariance matrix average} (CMA). In the HBA ap

Pratik Chakraborty, Aryan Bhargava

Proteins perform essential biological functions, and accurate classification of their sequences is critical for understanding structure-function relationships, enzyme mechanisms, and molecular interactions. This study presents a deep learning-based framework for functional group classification of protein sequences derived from the Protein Data Bank (PDB). Fo

Melika Ayoughi, Pascal Mettes, Paul Groth

Hyperbolic geometry is an effective geometry for embedding hierarchical data structures. Hyperbolic learning has therefore become increasingly prominent in machine learning applications where data is hierarchically organized or governed by hierarchical semantics, ranging from recommendation systems to computer vision. The quality of hyperbolic embeddings is

Hisham Sati, Urs Schreiber

Topological Chern phases of quantum materials, as well as brane charges on M-theory orbifolds, have famously been argued to be classified by (orbi) topological K-theory, or possibly by other stable and, notably, complex-oriented cohomology theories, such as elliptic cohomology or Morava K-theory. However, closer inspection reveals that the most fine-grained

M. Houzet, T. Vakhtel, J. S. Meyer

In a SQUID tuned away from half-integer flux (in units of the superconducting flux quantum), the concurrence of multiple Josephson harmonics and an asymmetry between the junctions leads to the Josephson diode effect -- a nonreciprocal current-voltage characteristic manifested as an asymmetry of critical currents at opposite polarities. We predict a dual vers

Yaniv Fogel, Meir Feder

We study batch learning with log-loss in the individual setting, where the outcome sequence is deterministic. Because empirical statistics are not directly applicable in this regime, obtaining regret guarantees for batch learning has long posed a fundamental challenge. We propose a natural criterion based on leave-one-out regret and analyze its minimax value

Joanna M. Rankin, Vyacheslav Zakharenko, Oleg Ulyanov, Ihor Kravtsov

PSR B1237+25 is perhaps the canonical example of a pulsar with a core/double cone profile. Moreover, it is bright with little spectral turnover, and its profile perhaps uniquely remains undistorted by scattering far into the decametric band. Here we assemble more than a dozen of the highest quality profiles (30 MHz to 5 GHz) from half a dozen observatories,

Sushrut Thorat, Adrien Doerig, Alexander Kroner, Carmen Amme

Scenes are complex, yet structured collections of parts, including objects and surfaces, that exhibit spatial and semantic relations to one another. An effective visual system therefore needs unified scene representations that relate scene parts to their location and their co-occurrence. We hypothesize that this structure can be learned self-supervised from

George Z. Li, Jason Li, Satish Rao, Junkai Zhang

Consider the single-source shortest paths problem on a directed graph with real-valued edge weights. We solve this problem in $O(n^{2.5}\log^{4.5}n)$ time, improving on prior work of Fineman (STOC 2024) and Huang-Jin-Quanrud (SODA 2025, 2026) on dense graphs. Our main technique is an shortcutting procedure that iteratively reduces the number of negative-weig

Parisa Omidvar, Marc Serra-Garcia

Structural nonlinearity can be harnessed to program complex functionalities in robotic devices. However, it remains a challenge to design nonlinear systems that will accomplish a specific, desired task. The responses that we typically describe as intelligent -- such a robot navigating a maze -- require a large number of degrees of freedom and cannot be captu

Pedro Ilídio, Felipe Kenji Nakano, Alireza Gharahighehi, Robbe D'hondt

Bipartite learning is a machine learning task that aims to predict interactions between pairs of instances. It has been applied to various domains, including drug-target interactions, RNA-disease associations, and regulatory network inference. Despite being widely investigated, current methods still present drawbacks, as they are often designed for a specifi

Christopher Cruz

Large language models (LLMs) are increasingly deployed in multi-turn dialogue settings, yet their behavior remains bottlenecked by naive history management strategies. Replaying the full conversation at every turn is simple but costly, while recency-based truncation or static summarization often causes early, high-impact user constraints to drift out of effe

Nelson Violante-Carvalho, Thiago de Paula, Leandro Calado, Felipe Marques dos Santos

Ocean eddy dipoles are among the most common mesoscale features and may be ubiquitous across the global oceans. However, wave-current interactions in their proximity have not been extensively studied. Here we examine the impact of surface currents on the wave field near an ocean eddy dipole. Using the WW3 wave model, we conducted idealized numerical simulati

Shalini Jangra, Zaid Almahmoud, Suparna De, Gareth Tyson

Reddit is in the minority of mainstream social platforms that permit posting content that may be considered to be at the edge of what is permissible, including so-called Not Safe For Work (NSFW) content. However, NSFW is becoming more common on mainstream platforms, with X now allowing such material. We examine the top 15 NSFW-restricted subreddits by size t

Sangwoo Seo, Hyunsung Kim, Jiwan Kim, Chanyoung Park

Mesh-based simulation using Graph Neural Networks (GNNs) has been recognized as a promising approach for modeling fluid dynamics. However, the mesh refinement techniques which allocate finer resolution to regions with steep gradients can induce the over-squashing problem in mesh-based GNNs, which prevents the capture of long-range physical interactions. Conv

Kaiser Hamid, Can Cui, Khandakar Ashrafi Akbar, Ziran Wang

Understanding not only where drivers look but also why their attention shifts is essential for interpretable human-AI collaboration in autonomous driving. Driver attention is not purely perceptual but semantically structured. Thus, attention shifts can be learned through minimal semantic supervision rather than dense large-scale annotation. We present \textb

Raven Gassis, Matthew B. Bayliss, Michael McDonald, Keren Sharon

We present the results of our dynamical state proxy measurements performed on 28 strong lensing galaxy clusters from the Sloan Giant Arcs Survey (SGAS). Using Chandra ACIS-I/S X-ray data supplemented with HST WFC3 imaging, we measure four morphological parameters: the concentration parameter (c), asymmetry parameter (A), centroid shift (log(w)), and the X-ra

Daniel Furelos-Blanco, Charles Pert, Frederik Kelbel, Alex F. Spies

Training general agents to follow complex instructions (tasks) in intricate environments (levels) remains a core challenge in reinforcement learning. Random sampling of task-level pairs often produces unsolvable combinations, highlighting the need to co-design tasks and levels. While unsupervised environment design (UED) has proven effective at automatically

Stephen Wright, Colin Paterson

Many mathematical modelling tasks (such as in Economics and Finance) are informed by data that is "found" rather than being the result of carefully designed experiments. This often results in data series that are short, noisy, multidimensional and contaminated with outliers, regime shifts, and confounding, uninformative or co-linear variables. We present a g

Herman Errico

Intentional threats are a major risk factor related to vulnerabilities in critical infrastructure assets, and an accurate risk assessment is necessary to analyze threats, assess vulnerabilities, and evaluate potential impacts on assets and systems. This research proposes a methodology that can be added as an additional phase in the risk assessment process. T

Qingqing Song, Shaoliang Xia

This paper introduces a concept for change-point detection based on normalized entropy as a fundamental metric, aiming to overcome the dependence of traditional entropy methods on assumptions about data distribution and absolute scales. Normalized entropy maps entropy values to the [0,1] interval through standardization, accurately capturing relative changes

Safaeid Hossain Arib, Rabeya Akter, Abdul Monaf Chowdhury, Md Jubair Ahmed Sourov

Object counting has achieved remarkable success on visible instances, yet state-of-the-art (SOTA) methods fail under occlusion. This failure stems from a fundamental architectural limitation where backbone networks encode occluding surfaces rather than target objects, thereby corrupting the feature representations required for accurate enumeration. To addres

Lukas Picek, César Leblanc, Alexis Joly, Pierre Bonnet

This paper describes GeoPl@ntNet, an interactive web application designed to make Essential Biodiversity Variables accessible and understandable to everyone through dynamic maps and fact sheets. Its core purpose is to allow users to explore high-resolution AI-generated maps of species distributions, habitat types, and biodiversity indicators across Europe. T

Matthew Duschenes, Diego García-Martín, Zoë Holmes, M. Cerezo

Moments of ensembles of unitaries play a central role in quantum information theory as they capture the statistical properties of dynamics of systems with some form of randomness. Indeed, concepts such as approximate $t$-designs arise when comparing how close an associated moment operator of a given unitary ensemble is to that of another, reference ensemble.

Siyaxolisa Kabane

We investigate the generalization properties of dense text embeddings when the embedding backbone is a large language model (LLM) versus when it is a non-LLM encoder, and we study the extent to which spherical linear interpolation (SLERP) model-merging mitigates over-specialization introduced by task-specific adaptation (e.g., LoRA). To make the comparison c

Kenichiro McAlinn, Kōsaku Takanashi

Cross-validation (CV) is routinely used across the sciences to select models and tune parameters, and the resulting choices are often interpreted as substantive scientific conclusions (e.g., which variables, mechanisms, or risk factors are ``supported by the data''). A key part of the CV procedure -- the hold-out size, or equivalently the fold count $K$ -- i

Soumendra Kumar Das, Prasanjit Samal, Brahmananda Chakraborty, Sridhar Sahu

Environmental oxidation has a great impact in tuning the physical, chemical and electronic properties of two-dimensional (2D) monolayers which can affect their practical applications in nanoscale engineering devices under ambient conditions. aza-triphenylene is a recently synthesized 2D materials whose practcal applications have not been systematically studi

E. Nikoghosyan, D. Baghdasaryan, D. Andreasyan, N. Azatyan

Research on young stellar populations is essential to understand the properties of embedded clusters and advance theories of their formation. This has driven advancements in methodologies for star detection, leading to the development of valuable databases and software. We present the scientific justification and operating principles of the IdentYS tool, whi

Minghui Chen, Hrad Ghoukasian, Ruinan Jin, Zehua Wang

Federated Learning (FL) enables decentralized, privacy-preserving model training but struggles to balance global generalization and local personalization due to non-identical data distributions across clients. Personalized Fine-Tuning (PFT), a popular post-hoc solution, fine-tunes the final global model locally but often overfits to skewed client distributio

Mohamed A. Mabrok, Yalda Zafari

State Space Models (SSMs), particularly the Mamba architecture, have recently emerged as powerful alternatives to Transformers for sequence modeling, offering linear computational complexity while achieving competitive performance. Yet, despite their effectiveness, understanding how these Vision SSMs process spatial information remains challenging due to the

Sushant Gautam, Michael A. Riegler, Pål Halvorsen

Vision-language models (VLMs) enable open-ended visual question answering but remain prone to hallucinations. We present HEDGE, a unified framework for hallucination detection that combines controlled visual perturbations, semantic clustering, and robust uncertainty metrics. HEDGE integrates sampling, distortion synthesis, clustering (entailment- and embeddi

Antonio Agresti, Max Sauerbrey, Mark Veraar

The celebrated De Giorgi-Nash-Moser theory ensures that solutions to uniformly elliptic or parabolic PDEs are bounded and H\"older continuous, even with merely bounded measurable coefficients. For parabolic SPDEs with transport noise, boundedness has recently been established, but H\"older continuity remains a key open problem in the regularity theory of par

Shuaike Shen, Ke Liu, Jiaqing Xie, Shangde Gao

Foundation models for medical image segmentation struggle under out-of-distribution (OOD) shifts, often producing fragmented false positives on OOD tumors. We introduce R$^{2}$Seg, a training-free framework for robust OOD tumor segmentation that operates via a two-stage Reason-and-Reject process. First, the Reason step employs an LLM-guided anatomical reason

Sina Rashidi, Hossein Sameti

Direct speech-to-speech translation (S2ST), in which all components are trained jointly, is an attractive alternative to cascaded systems because it offers a simpler pipeline and lower inference latency. However, direct S2ST models require large amounts of parallel speech data in the source and target languages, which are rarely available for low-resource la

Harshana Weligampola, Yuanrui Chen, Weiheng Tang, Qi Guo

Metalenses offer a path toward creating ultra-thin optical systems, but they inherently suffer from severe, spatially varying optical aberrations, especially chromatic aberration, which makes image reconstruction a significant challenge. This paper presents a novel algorithmic solution to this problem, designed to reconstruct a sharp, full-color image from t

Kaicheng Jin, Yang Peng, Jiansheng Yang, Zhihua Zhang

In this paper, we study the finite-sample statistical rates of distributional temporal difference (TD) learning with linear function approximation. The purpose of distributional TD learning is to estimate the return distribution of a discounted Markov decision process for a given policy. Previous works on statistical analysis of distributional TD learning fo

Partha S. Dey, Aditya S. Gopalan, Vijay G. Subramanian

We investigate the time to consensus in Nakamoto blockchains. Specifically, we consider two competing growth processes, labeled \emph{honest} and \emph{adversarial}, and determine the time after which the honest process permananetly exceeds the adversarial process. This is done via queueing techniques. The predominant difficulty is that the honest growth pro

Christian Berg

For an indeterminate Hamburger moment problem we consider an infinite family of analytic densities solving the moment problem and we prove that they all have finite (Shannon) entropy. These densities are either all bounded or all unbounded. The result is illustrated by the Al-Salam--Carlitz moment problem, where all the densities in the family are bounded.

M. A. Pellegrini, A. E. Zalesski

We prove that every finite simple group of Lie type $G$ can be generated by three regular unipotent elements. In certain cases we show that two regular unipotents are sufficient to generate $G$.

Amirpasha Hedayat, Karthik Duraisamy

Weather prediction is a quintessential problem involving the forecasting of a complex, nonlinear, and chaotic high-dimensional dynamical system. This work introduces an efficient reduced-order modeling (ROM) framework for short-range weather prediction and investigates fundamental questions in dimensionality reduction and reduced order modeling of such syste

Cheonsol Lee, Youngsang Jeong, Jeongyeol Shin, Huiju Kim

The stock market is inherently complex, with interdependent relationships among companies, sectors, and financial indicators. Traditional research has largely focused on time-series forecasting and single-company analysis, relying on numerical data for stock price prediction. While such approaches can provide short-term insights, they are limited in capturin

RuGway Wu, Maximilian Prüfer, Jörg Schmiedmayer

Matter-wave optics is often viewed as a linear analogue of photonics, where noninteracting particles are coherently split, diffracted, and recombined, and interference arises from single-particle coherence. In ultracold quantum gases, however, interactions are intrinsic and can rival or exceed kinetic and optical energy scales. This drives matter-wave optics