March 2026 arXiv papers — page 49

Fitsum Sileshi Beyene, Christopher L. Dancy

Optical character recognition (OCR) and document understanding systems increasingly rely on large vision and vision-language models, yet evaluation remains centered on modern, Western, and institutional documents. This emphasis masks system behavior in historical and marginalized archives, where layout, typography, and material degradation shape interpretati

Luca Mannella, Stefano Di Carlo, Alessandro Savino

Real-Time Operating Systems (RTOSes) play a crucial role in safety-critical domains, where deterministic and predictable task execution is essential. Yet they are increasingly exposed to ionizing radiation, which can compromise system dependability. To assess FreeRTOS under such conditions, we introduce KRONOS, a software-based, non-intrusive post-propagatio

Eric Shen

We study linear operators on a finite-dimensional space whose Kippenhahn curves consist of concentric circles centered at the origin. We say that such operators have Circularity property. One class of examples is rotationally invariant operators. To every operator with norm at most one, we associate an infinite sequence of partial isometries and study when C

Yuki Yamazaki, Shingo Kobayashi, Akira Furusaki

We propose Majorana-assisted nonlocal spin correlation as a manifestation of Majorana nonlocality in quasi-one-dimensional (1D) Kitaev spin liquids. Focusing on the flux-free sector of the Kitaev honeycomb model in a quasi-1D geometry, we uncover its topological nature and show that it hosts Majorana zero modes localized at both ends, which are stabilized by

Yue Li, David Mimno, Unso Eun Seo Jo

Translating natural language to SQL for data retrieval has become more accessible thanks to code generation LLMs. But how hard is it to generate SQL code? While databases can become unbounded in complexity, the complexity of queries is bounded by real life utility and human needs. With a sample of 376 databases, we show that SQL queries, as translations of n

Simon Seiderer, Andreas Geilen, Luan N. Sliwa, Linqiao Gan

Stimulated Brillouin-Mandelstam scattering offers exceptional capabilities for photonic signal processing, but current platforms demand performance trade-offs between long interaction lengths, high gain, low optical losses, and practical implementation. Here, we demonstrate a novel platform based on the reversible freezing of a carbon disulfide filled liquid

Gian Gentinetta, Friederike Metz, William Kirby, Giuseppe Carleo

The computation of thermal properties of quantum many-body systems is a central challenge in our understanding of quantum mechanics. We introduce the Quantum Finite Temperature Lanczos Method (QFTLM), which extends the finite-temperature Lanczos method to quantum computers by combining real-time quantum Krylov methods with efficient preparation of typical st

Guilherme Ilário Correr, John Goold, Marco Cattaneo

Local Operator Entanglement (LOE) has emerged an indicator of quantum chaos in many-body systems. Numerical studies have shown that, in chaotic systems, LOE grows linearly in time and displays a volume-law behavior at late times, scaling proportionally with the number of local degrees of freedom. Despite extensive numerical evidence, complemented by analytic

Xin Wu, Fei Teng, Xingwang Li, Bin Zheng

Accurately forecasting GPU workloads is essential for AI infrastructure, enabling efficient scheduling, resource allocation, and power management. Modern workloads are highly volatile, multiple periodicity, and heterogeneous, making them challenging for traditional predictors. We propose PRISM, a primitive-based compositional forecasting framework combining

M. Polovinkin, N. Rybin, D. Maksimov, F. Valiev

The design of efficient electrolysis devices for pure metal production requires accurate data on the properties of the melts used in the process. This work focuses on two key systems for calcium production: the molten Ca-Cu alloy and the CaCl$_2$-KCl electrolyte. High-temperature experiments are often expensive and time-consuming; however, we demonstrate tha

Thomas Walker, Anna L. Marchant, Elliot Bentine, Oliver Buchmueller

We present a high-flux source of cold strontium atoms based on a two-dimensional magneto-optical trap (2D MOT) and a Zeeman slower. We use the source to load a 3D MOT in a separate science chamber, observing a loading rate of $4 \times 10^{10}$ atoms/s -- to our knowledge, the highest reported loading flux for strontium. To characterise the vacuum pressure i

Chenjie Xie, Li You, Ruirong Chen, Gaoning He

Low-altitude communications can promote the integration of aerial and terrestrial wireless resources, expand network coverage, and enhance transmission quality, thereby empowering the development of sixth-generation (6G) mobile communications. As an enabler for low-altitude transmission, 3D channel fingerprints (3D-CF), also referred to as the 3D radio map o

Marcos Paulo Belançon, Rafaela Valcarenghi, Marcelo Sandrini, Brenno Greatti

Oxyfluorides containing up to 80 wt% recycled glass from end-of-life solar panels have been investigated. Reduced processing temperature and high transparency have shown that the material has potential for optical applications. In this work, cerium-doped samples were investigated. Spectroscopic study reveals the presence of Ce$^{3+}$, and luminescence from t

Georg Kordowich, Julian Oelhaf, Siming Bayer, Andreas Maier

While conventional power system protection isolates faulty components only after a fault has occurred, fault prediction approaches try to detect faults before they can cause significant damage. Although initial studies have demonstrated successful proofs of concept, development is hindered by scarce field data and ineffective feature selection. To address th

Hyeonjun An, Sihyun Kim, Chaerim Lim, Hyunjoon Kim

Multimodal Large Language Models (MLLMs) have achieved remarkable advances by integrating text, image, and audio understanding within a unified architecture. However, existing distributed training frameworks remain fundamentally data-blind: they parallelize computation without accounting for variations in input data characteristics. This data unawareness lea

Diyar Altinses, Andreas Schwung

Modern multimodal systems deployed in industrial and safety-critical environments must remain reliable under partial sensor failures, signal degradation, or cross-modal inconsistencies. This work introduces a mathematically grounded framework for fault-tolerant multimodal representation learning that unifies self-supervised anomaly detection and error correc

Yasong Dai, Zeeshan Hayder, David Ahmedt-Aristizabal, Hongdong Li

Recent diffusion and flow matching models have demonstrated strong capabilities in image generation and editing by progressively removing noise through iterative sampling. While this enables flexible inversion for semantic-preserving edits, few-step sampling regimes suffer from poor forward process approximation, leading to degraded editing quality. Existing

Lalita Na Nongkhai, Jingyun Wang, Adam Wynn, Takahiko Mendori

This paper introduces the design and development of a framework that integrates a large language model (LLM) with a retrieval-augmented generation (RAG) approach leveraging both a knowledge graph and user interaction history. The framework is incorporated into a previously developed adaptive learning support system to assess learners' code, generate form

Samuel Quirino

We prove that the algebra of invariants of a complete path algebra under the action of a homogeneous group of continuous algebra automorphisms is a complete path algebra and preserves finite or tame representation type.

Saswata Bose, Suvadeep Maiti, Shivam Kumar Sharma, Mythirayee S

Accurate sleep staging is essential for diagnosing OSA and hypopnea in stroke patients. Although PSG is reliable, it is costly, labor-intensive, and manually scored. While deep learning enables automated EEG-based sleep staging in healthy subjects, our analysis shows poor generalization to clinical populations with disrupted sleep. Using Grad-CAM interpretat

Tsuyoshi Miyatsu, Myung-Ki Cheoun, Koichi Saito

We investigate quarkyonic matter within a relativistic quark model by combining the dual quarkyonic picture with the quark-meson coupling (QMC) model. Using relativistic gaussian quark wavefunctions for the nucleon, we construct the quarkyonic QMC (QQMC) model and study the properties of symmetric nuclear matter and pure neutron matter. We find that the quar

Oliver H. E. Philcox, Guilherme L. Pimentel, Chen Yang

Multi-field models of inflation typically assume that interactions between particles can be treated perturbatively. Strongly-coupled models provide an intriguing alternative and may offer novel inflationary phenomenology. We study the "unparticle" scenario, where the inflaton is weakly mixed with a strongly-coupled sector, specified by a (gapless) co

Harry-Dean Kenchington Goldsmith, Nemanja Jovanovic, Anusha Pai Asnodkar, Yoo Jung Kim

Astrophotonics is central to the next generation of astronomical instrumentation, enabling compact photonic integrated circuits for both ground-based observatories and future space missions. Beam combination for nulling interferometry suppresses starlight, revealing exoplanets and companions. Two-waveguide photonic combiners rely on symmetric evanescent, inh

Siddharth Iyer

We study a problem of Douglass and Ono concerning the smallest integer $n$ such that the partition function $p(n)$ begins with a specified string of digits $f$ in base $b$. By employing an elementary discrepancy framework, we establish new upper bounds that significantly improve upon previous results of Luca.

Felipe Diaz

We embed the covariant, gauge-invariant gravitational radiation criteria of Fernández-Álvarez and Senovilla, based in terms of conformal geometry and the Bel-Robinson tensor, into the hydrodynamic framework of gauge/gravity duality. This construction uncovers a direct correspondence between bulk gravitational waves and dissipative processes in the boundary t

Carlos Barredo Martínez, Torben Skrzypek

We consider type IIB string theory on $\mathrm{AdS}_5\times S^5/\mathbb{Z}_{L}$ orbifold spaces with generic $L$. Recent localisation results in the dual 4d $\mathcal{N}=2$ circular quiver gauge theories provide us with strong coupling expansions of certain correlators involving twisted half-BPS operators. To leading order, these results have been matched to

Xinyu Xu, Kehan Cai, Yubai Shi, Peichen Zhong

We develop FIRE-Swap, a first-principles framework for sampling intrinsic compositional structures in complex perovskites with machine-learning interatomic potentials (MLIPs). Using both dedicated and universal MLIPs, we study the relaxor lead magnesium niobate (PMN) and the solid solutions lead zirconate titanate (PZT) and lead strontium titanate (PST). Acr

H. Denizli, A. Senol, M. Köksal

Rare decays of the Z boson provide a sensitive probe for physics beyond the Standard Model (SM). This study investigates the $e^{+}e^{-} \to Z \to ν\barνγ$ process within the context of the Tera-Z programmes at future colliders such as the FCC-ee and CEPC. The SM predicts a one-loop branching ratio of $7.16 \times 10^{-10}$ for $Z \to ν\barνγ$, a value four

Shuxin Lin, Rimi Banerjee, Zheyu Cheng, Kohei Kawabata

The surface states of certain topological phases can be linked to a quantum anomaly: the violation of a classical symmetry by a field theory via a non-conserved current. This has been generalized to the case of a non-Hermitian (NH) chiral anomaly affecting the surfaces states of an NH Weyl phase. Here, we show that the NH anomaly inflow is mediated by contin

Shadeeb Hossain, Natalie Sommer, Neda Adib

This paper presents a dynamic gamification architecture for an Extended Reality Artificial Intelligence virtual training environment designed to enhance STEM education through immersive adaptive, and kinesthetic learning. The proposed system can be introduced in four phases: Introduction Phase, Component Development Phase, Fault Introduction and Correction P

Ghazal Kaviani, Yavuz Yarici, Seulgi Kim, Mohit Prabhushankar

Daily Activity Recordings for Artificial Intelligence (DARai, pronounced "Dahr-ree") is a multimodal, hierarchically annotated dataset constructed to understand human activities in real-world settings. DARai consists of continuous scripted and unscripted recordings of 50 participants in 10 different environments, totaling over 200 hours of data from

Neha Goregaokar, Aaron Lin

We consider real hyperplane arrangements whose hyperplanes are of the form $\{x_i - x_j = s\}$ for some integer $s$, which we call deformations of the braid arrangement. In 2018, Bernardi gave a counting formula for the number of regions of any deformation of the braid arrangement $\mathcal{A}$ as a signed sum over some decorated trees. He further showed tha

Trevor Karn

We provide simple presentations in terms of generators and relations for the invariant subring of both the Orlik--Solomon algebra and Varchenko--Gel'fand ring of the type $A_n$ reflection arrangement acted upon by the type $A_{n-1}$ reflection group. This may be interpreted as a presentation for the cohomology of the ``mixed configuration space" of $n$ red p

Kalle Kujanpää, Yuying Zhu, Kristina Klinkner, Shervin Malmasi

We investigate machine learning approaches for optimizing real-time staffing decisions in semi-automated warehouse sortation systems. Operational decision-making can be supported at different levels of abstraction, with different trade-offs. We evaluate two approaches, each in a matching simulation environment. First, we train custom Transformer-based polici

David Torres Ramos, Vihan Lakshman, Chen Luo, Todd Treangen

We study the problem of building space-efficient, in-memory indexes for massive key-value datasets with highly skewed value distributions. This challenge arises in many data-intensive domains and is particularly acute in computational genomics, where $k$-mer count tables can contain billions of entries dominated by a single frequent value. While recent work

Rui Li, Xing Zhang, Qiming Sun, Yuanheng Wang

We introduce a GPU-accelerated multigrid Gaussian-Plane-Wave density fitting (FFTDF) approach for efficient Fock builds and nuclear gradient evaluations within Kohn-Sham density functional theory, as implemented in the GPU4PySCF module of PySCF. Our CUDA kernels employ a grid-based parallelization strategy for contracting Gaussian basis function pairs and ac

Yuta Inoue, Ken-ichi Kawarabayashi, Atsuyuki Miyashita, Bojan Mohar

We give a near-linear time 4-coloring algorithm for planar graphs, improving on the previous quadratic time algorithm by Robertson et al. from 1996. Such an algorithm cannot be achieved by the known proofs of the Four Color Theorem (4CT). Technically speaking, we show the following significant generalization of the 4CT: every planar triangulation contains li

Pedro Oliveira, Tayana Conte, Marco Gerosa, Igor Steinmacher

Open source software (OSS) sustainability depends not only on code contributions but also on governance structures that define who decides, who acts, and how responsibility is distributed. We lack systematic empirical evidence of how projects formally codify roles and authority in written artifacts. This paper investigates how OSS projects define and structu

Jiasun Li, Project Team

The growing replication crisis across disciplines such as economics, finance, and other social sciences as well as computer science undermines the credibility of academic research. Current institutional solutions -- such as artifact evaluations and replication packages -- suffer from critical limitations, including shortages of qualified data editors, diffic

Venkatesh Sivaraman, Patrick Vossler, Adam Perer, Julian Hong

Generative artificial intelligence (AI) tools can now help people perform complex data science tasks regardless of their expertise. While these tools have great potential to help more people work with data, their end-to-end approach does not support users in evaluating alternative approaches and reformulating problems, both critical to solving open-ended tas

Xiaoyu Tang, Jun Dong, Jintao Cheng, Rui Fan

Remote sensing visual grounding (RSVG) aims to localize specific targets in remote sensing images using natural language expressions. However, existing methods are restricted to single-sensor domains, i.e., either optical or synthetic aperture radar (SAR), limiting their real-world applicability. In this paper, we introduce the Cross-Domain RSVG (CD-RSVG) ta

Qinyan Shen, Karl Gregory, Xianzheng Huang

We propose a unified framework to draw inferences for regression coefficients in a generalized linear model (GLM) following Lasso-based variable selection. We adapt to non-Gaussian GLMs a recently developed parametric programming strategy for post-selection inference in the linear model with a Gaussian response by drawing parallels between maximum likelihood

Nicolò Lo Piparo, William J. Munro, Kae Nemoto

We investigate congestion-aware control of quantum repeater nodes operating under stochastic traffic and finite memory coherence. Entanglement generation is modeled as a probabilistic process producing Werner states subject to depolarizing memory decoherence, while entanglement requests arrive according to Poisson and bursty ON--OFF processes. Using a queuei

Hongze Gao, Da Zhou, Nguyen Tuan Hung, Chengdong Wang

The recent experimental realization of two-dimensional (2D) transition metal nitrides (TMNs, e.g., Mo5N6, {\delta}-MoN, and W5N6) opens new opportunities for exploring their fundamental physical properties at the two-dimensional limit. In this work, we propose a unified picture of transport phenomena in the nitrogen-rich 2D W5N6 and Mo5N6, and the stoichiome

Bruce W. Brewer, Haitao Wang

Let $S$ be a set of $n$ points in a polygon $P$ with $m$ vertices. The geodesic unit-disk graph $G(S)$ induced by $S$ has vertex set $S$ and contains an edge between two vertices whenever their geodesic distance in $P$ is at most one. In the weighted version, each edge is assigned weight equal to the geodesic distance between its endpoints; in the unweighted

Basil Grammaticos, Alfred Ramani, Ralph Willox

We examine the Lyness mapping (an integrable $N$th-order discrete system which can be generated from a one-dimensional reduction of the Hirota-Miwa equation) from the point of view of deautonomisation. We show that only the $N=2$ case can be deautonomised when one works with the standard form of the mapping. However it turns out that deautonomisation is poss

Bryce M. Barclay, Alex Mahalov

Fundamental to the analysis of nonlinear relativistic motion is the precise characterization of the induced dynamic Doppler effects. In this work, we analyze the electromagnetic signals observed by non-inertial receivers using two frameworks to describe the relativistic motion. We first consider observer paths described by higher-order kinematic 4 vectors: r

Xiaolong Hans Han, Ruojing Jiang

We consider closed hypersurfaces smoothly immersed in hyperbolic manifolds up to homotopy and commensurability. We prove that if a closed hyperbolic manifold $M$ contains a sequence of asymptotically geodesic hypersurfaces, then $\pi_1(M)$ is virtually special and hence linear over integers. If $M$ (dimension at least 3) is, in addition, arithmetic of type I

S. P. Kish, H. J. Vallury, J. Pieprzyk, C. Thapa

We introduce Quantum Spectral Authentication (QSA), a symmetric-key entity-authentication and key-derivation protocol in which a remote endpoint proves it still holds a hidden planted state, an eigenstate of the challenge it can prepare, without revealing it. Each round issues a fresh public unitary challenge with its own planted state, and the endpoint retu

Soubhik Biswas, Andrew Dowdy, Savin Chand

Understanding how weather and climate influence fire risk is important for many purposes, including climate adaptation planning and decision-making in sectors such as emergency management, finance, health and infrastructure (e.g., for energy and water availability). In this study, bias-corrected 20CRv2c reanalysis data are used to investigate the climatology

Luyu Yang, Yutong Dai, An Yan, Viraj Prabhu

The physical world is not merely visual; it is governed by rigorous structural and procedural constraints. Yet, the evaluation of vision-language models (VLMs) remains heavily skewed toward perceptual realism, prioritizing the generation of visually plausible 3D layouts, shapes, and appearances. Current benchmarks rarely test whether models grasp the step-by

Shanxia Wang

We introduce a dual-threshold probabilistic knowing value logic for uncertain multi-agent settings. The framework captures within a single formalism both probabilistic-threshold attitudes toward propositions and high-confidence attitudes toward term values, thereby connecting probabilistic epistemic logic with classical knowing value logic. It is especially

Daniel Pierce, Renuka Rajapakse

An electron in quantum confinement takes on a discrete energy spectrum which is defined based on the solution to the Schrodinger Equation for a given potential. Well defined closed-form energy spectra are known for the particle in a box, circular potential, quarter circle potential, and an equilateral triangle. A closed-form solution for more complex shapes

Alejandro González I., Pedro A. Orellana, Vladimir Juricic

We investigate bound states in the continuum (BICs) in a hybrid normal--superconducting triple quantum dot system, where the central dot is coupled to two normal leads and the lateral dots are proximity-coupled to superconducting electrodes. Local electron--electron interactions are treated within the Hubbard approximation. Finite bias, together with lateral

Joel Cortez Osuna, Sarah Shandera

While primordial black holes (PBHs) have long been a benchmark target for microlensing searches, the modern landscape of dark matter models suggests other, distinct, formation channels for compact objects made of dark matter. In the large class of self-interacting, dissipative models, dark matter has cooling channels that can enable fragmentation and gravita

Zhuoran Li, Hanieh Totonchi Asl, Yifei Cai, Ebrahim Nouri

Secure multi-party computation (MPC) offers a practical foundation for privacy-preserving machine learning at the edge. However, current MPC systems rely heavily on communication and computation-intensive primitives-such as secure comparison for nonlinear inference, which are often impractical on resource-constrained platforms. To enable real-time inference

Jörn Stöhler, Stefan Blügel, Christoph Friedrich

We describe an all-electron implementation of the Bethe-Salpeter equation (BSE) for the calculation of optical absorption spectra in the full-potential linearized augmented-plane-wave (FLAPW) method. So far, FLAPW implementations have resorted to a simple plane-wave basis for the bare and screened Coulomb potentials, thereby forgoing the all-electron descrip

Kai Z. Teh, Kayvan Sadeghi, Terry Soo

In applications, quantities of interest are often modelled in equilibrium or an equilibrium solution is sought. The presence of confounding makes causal inference in this setting challenging. We provide interpretable graphical models for equilibrium systems with confounding using anterial graphs (Lauritzen and Sadeghi, 2018), a class of graphs containing dir

Zhenyi Wang, Siyu Luan

As machine learning (ML) systems expand in both scale and functionality, the security landscape has become increasingly complex, with a proliferation of attacks and defenses. However, existing studies largely treat these threats in isolation, lacking a coherent framework to expose their shared principles and interdependencies. This fragmented view hinders sy

Kliment Ho, Ilya Zaslavsky

Emergency response systems generate data from many agencies and systems. In practice, correlating and updating this information across sources in a way that aligns with Next Generation 9-1-1 data standards remains challenging. Ideally, this data should be treated as a continuous stream of operational updates, where new facts are integrated immediately to pro

Muhammad Mahmudul Hasan, Ingrid Torres, Alex Krasnok

Damage in infrastructure is often hidden until it becomes costly or dangerous. Common examples include corrosion under insulation, early fatigue damage in steel, corrosion of embedded reinforcement, and abnormal current flow in batteries and power equipment. Magnetic methods are attractive because they can sense through coatings, insulation, and concrete cov

Bernhard Vogginger, Vasilis Thanasoulis, Johannes Partzsch, Christian Mayr

Neuromorphic VLSI systems take inspiration from biology to enable efficient emulation of large-scale spiking neural networks and to explore new computational paradigms. To establish large neuromorphic systems, a sophisticated routing infrastructure is needed to communicate spikes between chips and to/from the host computer. For the BrainScaleS wafer-scale ne

Silvia Rossi, Diletta Huyskes, Mackenzie Jorgensen

Ethical debates in AI have primarily focused on back-end issues such as data governance, model training, and algorithmic decision-making. Less attention has been paid to the ethical significance of front-end design choices, such as the interaction and representation-based elements through which users interact with AI systems. This gap is particularly signifi

Montie Avery, Paul Carter, Björn de Rijk

We establish nonlinear stability of fronts that describe the creation of a periodic pattern through the invasion of an unstable state. Our results concern pushed fronts, that is, fronts whose propagation is driven by a localized mode at the front interface. We prove that these pushed pattern-forming fronts attract initial data supported on a half-line, and t

Lukas Kratochvila, Jakub Stefansky, Simon Bilik, Robert Rous

Fire safety consists of a complex pipeline, and it is a very important topic of concern. One of its frontal parts are the smoke detectors, which are supposed to provide an alarm prior to a massive fire appears. As they are often difficult to reach due to high ceilings or problematic locations, an automatic inspection system would be very beneficial as it cou

Nikolas Papadopoulos, Shreenithi Navaneethan, Sheng Bai, Ankur Samanta

Preference learning methods, such as Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO), rely on pairwise human judgments, yet little is known about the cognitive processes underlying these judgments. We investigate whether eye-tracking can reveal preference formation during pairwise AI-generated image evaluation. Thir

Víctor M. Rivilla, Miguel Sanz-Novo, David San Andrés

The increasing detection of new molecules in the interstellar medium (ISM) shows that stereoisomerism is a fundamental contributor to interstellar molecular complexity. This work presents the first comprehensive overview of interstellar stereoisomerism. A total of 16 stereoisomeric pairs have been identified (13 conformational and 3 geometric), spanning mole

Katarina Trojachanec Dineva, Stefan Andonov, Ilinka Ivanoska, Ivan Kitanovski

Recent advances in multimodal large language models enable new possibilities for image-based decision support. However, their reliability and operational trade-offs in neuroimaging remain insufficiently understood. We present a comprehensive benchmarking study of vision-enabled large language models for 2D neuroimaging using curated MRI and CT datasets cover

John M. Campbell, Yuka Yamaguchi

Recently, the second author [Ramanujan J. 2026] introduced and proved a $q$-series identity that appears to provide the first known $q$-analogue of an evaluation for a ${}_{2}F_{1}$-series known as \emph{Gosper's strange series}. Yamaguchi's derivation of this $q$-analogue relies on three-term relations for ${}_{2}\phi_{1}$-series along with Heine's transfor

Jelena Markovic-Voronov, Kayhan Behdin, Yuanda Xu, Zhengze Zhou

We study the problem of routing queries to large language models (LLMs) under cost, GPU resources, and concurrency constraints. Prior per-query routing methods often fail to control batch-level cost, especially under non-uniform or adversarial batching. To address this, we propose a batch-level, resource-aware routing framework that jointly optimizes model a

Robert Reinertsen

We present a simulation-based evaluation of the Inference Headroom Ratio (IHR), a dimensionless diagnostic quantity for characterizing inference stability in constrained decision systems. IHR formalizes the relationship between a system's effective inferential capacity C and the combined uncertainty and constraint load U + K imposed by its operating environm

Isha Puri, Mehul Damani, Idan Shenfeld, Marzyeh Ghassemi

Given a question, a language model (LM) implicitly encodes a distribution over possible answers. In practice, post-training procedures for LMs often collapse this distribution onto a single dominant mode. While this is generally not a problem for benchmark-style evaluations that assume one correct answer, many real-world tasks inherently involve multiple val

Dipendu Halder, Srijata Lahiri, Saurabh Basu

Non-Hermitian skin effect, which is a unique feature of non-Hermitian systems, exhibits the formation of an extensive number of boundary modes under open boundary conditions. However, its manifestation in higher dimensions remains elusive. In our work, we demonstrate a hybrid skin-topological effect arising from the interplay between first-order band topolog

Shrey Lingampalli

The institutionalization of stablecoins has led to a paradigm shift in reserve management, accelerated by the 2025 Green Energy and National Infrastructure Underpinning Stablecoins (GENIUS) Act. This study investigates the "Climate-Liquidity Nexus," defined as the structural vulnerability arising from the use of environmentally sustainable but secondary-mark

Nathan Strange

Spacecraft development costs remain high despite falling launch costs, in part because Model-Based Systems Engineering (MBSE) tools carry the complexity of the object-oriented programming paradigm: tightly coupled data and logic, mutable state, and rigid class hierarchies that resist integration with discipline-specific analysis tools. This paper presents a

Haobo Xu, Sirui Chen, Ruizhong Qiu, Yuchen Yan

Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced the reasoning capabilities of Large Language Models (LLMs). However, methods such as GRPO and DAPO suffer from substantial computational cost, since they rely on sampling many rollouts for each prompt. Moreover, in RLVR the relative advantage is often sparse: many samples become

Alan B. H. Nguyen, Gregory Walth, Ashley J. Ross, James W. Colbert

We consider the application of a ubercalibration-like relative flux calibration to the grism observations of the Roman High Latitude Wide Area Survey (HLWAS). We propose a simplified model of the calibration with an independent flat field for each detector in each exposure of the focal plane. In addition, we include two wavelength dependent components: a sin

Hayley J. Macpherson

In this work we investigate the weak lensing convergence using an end-to-end nonlinear general relativistic framework. Combining numerical relativity simulations of large-scale structure formation with general relativistic ray-tracing, we compare our nonlinear calculation to the expectation based on perturbation theory for a set of 20 synthetic observers. We

Ahmed Lekssays

Large Language Models (LLMs) face critical challenges when analyzing security vulnerabilities in real world codebases: token limits prevent loading entire repositories, code embeddings fail to capture inter procedural data flows, and LLMs struggle to generate complex static analysis queries. These limitations force existing approaches to operate on isolated

Junyi Ouyang, Wenbin Teng, Gonglin Chen, Yajie Zhao

Long-trajectory video generation is a crucial yet challenging task for world modeling primarily due to the limited scalability of existing video diffusion models (VDMs). Autoregressive models, while offering infinite rollout, suffer from visual drift and poor controllability. To address these issues, we propose DCARL, a novel divide-and-conquer, autoregressi

Ana Petrovic, Maarten van Ham, David Manley, Tiit Tammaru

There are relatively few comparative cross-European studies on segregation, and those that do exist often use a single measure of segregation at a single spatial scale. This paper investigates ethnic segregation in seven European capitals (Amsterdam, Berlin, Lisbon, London, Madrid, Paris, and Rome) using the five dimensions of segregation (centralisation, ev

Yong Xie, Kexin He, Andres Castellanos-Gomez

The control of complex laboratory instrumentation often requires significant programming expertise, creating a barrier for researchers lacking computational skills. This work explores the potential of large language models (LLMs), such as ChatGPT, and LLM-based artificial intelligence (AI) agents to enable efficient programming and automation of scientific e

Vineet Kumar Sharma, Alana Okullo, Barun Ghosh, Arun Bansil

Paucity of naturally occurring noncentrosymmetric materials is stimulating growing interest in engineered two-dimensional systems for nonlinear optical applications. Here, we show that breaking inversion symmetry in centrosymmetric bilayer Bi$_2$Se$_3$ through twisting, point-defect insertion, or the application of an external electric field unlocks rich non

Anish Agarwal, Jungjun Choi, Ming Yuan

We introduce a flexible framework for high-dimensional matrix estimation to incorporate side information for both rows and columns. Existing approaches, such as inductive matrix completion, often impose restrictive structure-for example, an exact low-rank covariate interaction term, linear covariate effects, and limited ability to exploit components explaine

Polina Matveeva, Dmitri Gutman, Sam T. Carr

We study topology in gapless phases of an interacting spinful model with spin-charge separation. We focus on the gapless boundaries between $\mathbb{Z}_2$ symmetry-breaking phases. We find two topologically non-trivial gapless states that occur at the boundary between a non-trivial and a trivial insulator. They correspond to topological Luther-Emery liquid a

Tom Bullock, Emily Machniak, You-Jin Kim, Radha Kumaran

Tracking moving objects is a critical skill for many everyday tasks, such as crossing a busy street, driving a car or catching a ball. Attention is a key cognitive function that supports object tracking; however, our understanding of the brain mechanisms that support attention is almost exclusively based on evidence from tasks that present stable objects at

Francesco Ruscelli

We propose a general framework to extend Flow Matching to homogeneous spaces, i.e. quotients of Lie groups. Our approach reformulates the problem as a flow matching task on the underlying Lie group by lifting the data distributions. This strategy avoids the potentially complicated geometry of homogeneous spaces by working directly on Lie groups, which in tur

Yongda Fan, John Wu, Andrea Fitzpatrick, Naveen Baskaran

Clinical decisions are high-stakes and require explicit justification, making model interpretability essential for auditing deep clinical models prior to deployment. As the ecosystem of model architectures and explainability methods expands, critical questions remain: Do architectural features like attention improve explainability? Do interpretability approa

Harrison Li, Kevin Wang, Cheol Jun Cho, Jiachen Lian

Building a diagnosis model for primary progressive aphasia (PPA) has been challenging due to the data scarcity. Collecting clinical data at scale is limited by the high vulnerability of clinical population and the high cost of expert labeling. To circumvent this, previous studies simulate dysfluent speech to generate training data. However, those approaches

Thales Sales Almeida, Rodrigo Nogueira, Hélio Pedrini

Synthetic data generation through document rewriting has emerged as a promising technique for improving language model pretraining, yet most studies focus on English and do not systematically control for the quality of the source data being rewritten. We present a controlled study of how synthetic rewriting interacts with source data quality in the context o

Chih-En Lin, Attreyee Mukherjee, Ajay Rawat, Ruqi Zhang

Patch reviewing is critical for software development, especially in distributed open-source development, which highly depends on voluntary work, such as Linux. This paper studies the past 10 years of patch reviews of the Linux memory management subsystem to characterize the challenges involved in patch reviewing at scale. Our study reveals that the review pr

Fedor B. Lyudogovskiy

We study the outer geometry of the partition graph $G_n$, focusing on its canonical front-and-side framework, the family of nontrivial rectangular partitions, and the rear structures suggested by the visible geometry of the graph. We formalize the boundary framework $\mathcal B_n=\mathcal M_n\cup\mathcal L_n\cup\mathcal R_n$, where $\mathcal M_n$ is the main

Michail Karatarakis, Freek Wiedijk

We formalize Hilbert's Seventh Problem and its solution, the Gelfond-Schneider theorem, in the Lean 4 proof assistant. The theorem states that if $\alpha$ and $\beta$ are algebraic numbers with $\alpha \neq 0,1$ and $\beta$ irrational, then $\alpha^\beta$ is transcendental. Originally proven independently by Gelfond and Schneider in 1934, this result is a co

Younes Javanmard

We introduce a coefficient-decoupled matrix product operator (MPO) representation for Pauli-sum operators that separates reusable symbolic operator support from a tunable coefficient bridge across a fixed bipartition. This representation provides a direct interface to linear-combination-of-unitaries (LCU) circuits: the symbolic left/right dictionaries define

Alabi Mehzabin Anisha, Guangjing Wang, Sriram Chellappan

State-of-the-art crowd counting and localization are primarily modeled using two paradigms: density maps and point regression. Given the field's security ramifications, there is active interest in model robustness against adversarial attacks. Recent studies have demonstrated transferability across density-map-based approaches via adversarial patches, but cro

Hayder Saad Abdulbaqi, Mohammed Hadi Rahim, Mohammed Hassan Hadi, Haider Ali Aboud

MRI-based medical imaging has become indispensable in modern clinical diagnosis, particularly for brain tumor detection. However, the rapid growth in data volume poses challenges for conventional diagnostic approaches. Although deep learning has shown strong performance in automated classification, many existing solutions are confined to closed technical arc

Jean Ruppenthal

It is well known that the Grauert-Riemenschneider canonical sheaf $\mathcal{K}_X$ of holomorphic square-integrable $n$-forms is a central tool in $L^2$-theory for the $\overline\partial$-operator on a singular complex space $X$ of pure dimension $n$. It was shown a few years ago that a comprehensive $L^2$-theory requires also the study of the sheaf $\mathcal

Iqra Ali, Talia Tseriotou, Mahmud Elahi Akhter, Yuxiang Zhou

Longitudinal NLP tasks such as mental health monitoring and stance evolution require modeling temporally ordered text to track persistence and detect change. Such tasks also suffer from data scarcity, often involving rare events and sparsely annotated data. Large Language Models (LLMs) can learn from small amounts of data through in-context learning (ICL), w

Tanja Eisner, Valentin Gillet

We study the weak limit semigroup of an operator $T$, i.e., the set of all operators being weak limit points of the powers of $T$, in three different but related contexts: Koopman operators of measure-preserving transformations, contractions/isometries/unitaries on separable Hilbert spaces and positive operators on $L^p$-spaces. Hereby we focus on finding la

Yubo Wang, Marie Fridberg, Anirejuoritse Bafor, Ole Rahbek

Pin sites represent the interface where a metal pin or wire from the external environment passes through the skin into the internal environment of the limb. These pins or wires connect an external fixator to the bone to stabilize the bone segments in a patient with trauma or deformity. Because these pin sites represent an opportunity for external skin flora