October 2025 arXiv papers — page 129

Sam B. Ponnada, Philip F. Hopkins, Yue Samuel Lu, Emily M. Silich

Many state-of-the-art galaxy simulations featuring traditional feedback modes have significant challenges producing enough extended soft X-ray ($\sim 0.5-2$ keV) emission at R $\sim 0.5-1$ R$_{\rm vir}$ observed around galaxies with stellar masses M$_{\rm \ast} \lesssim 10^{11} \rm M_\odot$, without violating galaxy mass function constraints. Moreover, therm

Loïc Honet, Lorenzo Küchler, Adam Pound, Geoffrey Compère

With the upcoming third-generation gravitational-wave detectors comes the need to build complete, faithful, and fast waveform models for asymmetric-mass-ratio compact binaries. Most efforts within the self-force community have focused on modeling these binaries' inspiral regime, but for ground-based detectors the systems' final merger can represent the domin

Anson Hook, Junwu Huang, Mohamad Shalaby

We revisit and invalidate all dark photon dark matter constraints from resonant conversion of dark photons into photons (plasmons) in the early universe. These constraints rely on the resonant transfer of a substantial portion of the dark photon energy density into the SM plasma, heating the plasma in the process. We demonstrate that this resonant transfer s

Vishal Tiwari, Chi-Ho Chan, Tamara Bogdanović, Yan-Fei Jiang

We present a global three-dimensional radiation magnetohydrodynamic (RMHD) simulation of a circumbinary disk (CBD) around a massive black hole binary (MBHB) with a total mass $2 \times 10^7\,M_{\odot}$ and mass ratio $0.1$, separated by $100\, GM_{\rm tot}/c^2$. The inclusion of radiation makes the disk thinner, denser, less eccentric at the inner edge, and

William McClymont, Aaron Smith, Sandro Tacchella

Extragalactic nebular emission has long been a workhorse probe of the processes driving galaxy evolution, but the richness of JWST spectroscopy has shifted the bottleneck from data acquisition to physical interpretation and modelling. In this context, we present a major update to the Monte Carlo radiative transfer code COLT to facilitate self-consistent mode

Jiayin Gu, Shi-Jia Lin, Ding Yu Shao, Lian-Tao Wang

The unification of quantum information science and collider physics is opening a new frontier in high-energy experiments, making a systematic understanding of decoherence a critical challenge. We present a framework to systematically compute spin decoherence from final-state radiation by combining soft-collinear effective theory and open quantum system techn

Hugo Lóio, Guglielmo Lami, Lorenzo Leone, Max McGinley

We investigate how non-stabilizer resources enable the emergence of quantum state designs within the projected ensemble. Starting from initial states with finite magic and applying resource-free Clifford circuits to scramble them, we analyze the ensemble generated by performing projective Pauli measurements on a subsystem of the final state. Using both analy

Kazutaka Kimura, Kazuyuki Sugimura, Takashi Hosokawa, Hajime Fukushima

We present a radiation-hydrodynamics (RHD) scheme that enables 3D simulations resolving both protostellar interiors and their surrounding accretion flows within a single framework, to clarify how a protostar evolves while interacting with the accretion flow. The method builds on an explicit two-moment M1 closure scheme with a reduced speed of light approxima

Yang Bai, Keping Xie, Bei Zhou

We propose using current and future large-volume neutrino telescopes as ``Large Neutrino Colliders" (L$ν$Cs) to explore TeV-scale physics beyond the Standard Model. Cosmic neutrinos with energies above 100 PeV colliding with nucleons in the detector reach center-of-mass energies beyond the 14 TeV limit of the Large Hadron Collider (LHC). Using recently p

Rei Nishiura, Tsuyoshi Inoue

We investigate magnetic-field amplification driven by the nonresonant hybrid (NRH or Bell) instability and its impact on cosmic-ray (CR) acceleration at reverse shocks of ultrafast outflows (UFOs) from active galactic nuclei (AGN). Previous kinetic studies by particle-in-cell simulations have demonstrated that when maximum CR energy is near the injection sca

Johanna Borissova, João Magueijo

We consider the path-integral quantization of a minisuperspace cut-and-paste Lorentzian wormhole connecting two Minkowski spacetimes. The dynamics of the throat radius as a function of proper time is governed by a non-polynomial effective action derived by an application of the Israel junction condition formalism. Within a saddle-point approximation of the p

Andreas Kirchner, Berndt Müller, Jyotirmoy Roy, Chathuranga Sirimanna

An effective field theory framework is developed to study the interaction of heavy quarks in strongly coupled quark-gluon plasma (QGP). The latter is treated as a relativistic non-dissipative colorless fluid which can be studied using a derivatively coupled effective field theory based on previous work. Coupling this to heavy quarks provides a systematic way

Vijay Balasubramanian, Charlie Cummings

Tensor networks prepare states that share many features of states in quantum gravity. However, standard constructions are not diffeomorphism invariant and do not support an algebra of non-commuting area operators. Recently, analogues of both problems were addressed in a tensor network discretization of topological field theories (TFT) with finite or compact

Sihui Ji, Xi Chen, Xin Tao, Pengfei Wan

Video generation models nowadays are capable of generating visually realistic videos, but often fail to adhere to physical laws, limiting their ability to generate physically plausible videos and serve as ''world models''. To address this issue, we propose PhysMaster, which captures physical knowledge as a representation for guiding video generation models t

Anton Simen, Carlos Flores-Garrigós, Murilo Henrique De Oliveira, Gabriel Dario Alvarado Barrios

We introduce a Hamiltonian-based quantum feature extraction method that generates complex features via the dynamics of $k$-local many-body spins Hamiltonians, enhancing machine learning performance. Classical feature vectors are embedded into spin-glass Hamiltonians, where both single-variable contributions and higher-order correlations are represented throu

Apekshya Ghimire, Chandralekha Singh

In this research, we investigated the impact of peer collaboration and changes from individual to group performance of graduate students on the Conceptual Survey of Electricity and Magnetism (CSEM) without any guidance from the instructor. We define construction of knowledge as a case in which the group answered the question correctly but in the individual a

Xinchen Zhang, Xiaoying Zhang, Youbin Wu, Yanbin Cao

We introduce Generative Universal Verifier, a novel concept and plugin designed for next-generation multimodal reasoning in vision-language models and unified multimodal models, providing the fundamental capability of reflection and refinement on visual outcomes during the reasoning and generation process. This work makes three main contributions: (1) We bui

Xinhang Liu, Yuxi Xiao, Donny Y. Chen, Jiashi Feng

Effective spatio-temporal representation is fundamental to modeling, understanding, and predicting dynamics in videos. The atomic unit of a video, the pixel, traces a continuous 3D trajectory over time, serving as the primitive element of dynamics. Based on this principle, we propose representing any video as a Trajectory Field: a dense mapping that assigns

Alejandro Gil-García, C. S. Shahbazi

We characterize, in every dimension and signature, the algebraic squares of an irreducible complex spinor as a pair of exterior forms satisfying a prescribed system of algebraic relations that we present in terms of the geometric product of the underlying quadratic vector space. As a result, we obtain a general correspondence between irreducible complex spin

Yiming Chen, Zekun Qi, Wenyao Zhang, Xin Jin

In this paper, we claim that 3D visual grounding is the cornerstone of spatial reasoning and introduce the Grounded-Spatial Reasoner (GS-Reasoner) to explore the effective spatial representations that bridge the gap between them. Existing 3D LLMs suffer from the absence of a unified 3D representation capable of jointly capturing semantic and geometric inform

Jia-Chen Gu, Junyi Zhang, Di Wu, Yuankai Li

As retrieval-augmented generation (RAG) tackles complex tasks, increasingly expanded contexts offer richer information, but at the cost of higher latency and increased cognitive load on the model. To mitigate this bottleneck, especially for intricate multi-hop questions, we introduce BRIEF-Pro. It is a universal, lightweight compressor that distills relevant

Simon Corrodi, Mackenzie Devilbiss, E. Craig Dukes, Ralf Ehrlich

The cosmic ray veto (CRV) detector of the Mu2e experiment consists of four layers of plastic scintillation counters that surround the detector solenoid. These counters are embedded with wavelength-shifting fibers and are read out by silicon photomultipliers (SiPMs). The performance of a subset of the CRV counters was studied in a cosmic-ray test stand. Using

Giovanni Monea, Yair Feldman, Shankar Padmanabhan, Kianté Brantley

The scalability of large language models for long-context reasoning is severely constrained by the linear growth of their Transformer key-value cache, which incurs significant memory and computational costs. We posit that as a model generates reasoning tokens, the informational value of past generated tokens diminishes, creating an opportunity for compressio

Yi Zhang, Bolin Ni, Xin-Sheng Chen, Heng-Rui Zhang

Fully open multimodal large language models (MLLMs) currently lag behind proprietary counterparts, primarily due to a significant gap in data quality for supervised fine-tuning (SFT). Existing open-source datasets are often plagued by widespread noise and a critical deficit in complex reasoning data, such as Chain-of-Thought (CoT), which hinders the developm

Tuhin Chakrabarty, Jane C. Ginsburg, Paramveer Dhillon

The use of copyrighted books for training AI has sparked lawsuits from authors concerned about AI generating derivative content. Yet whether these models can produce high-quality literary text emulating authors' voices remains unclear. We conducted a preregistered study comparing MFA-trained writers with three frontier models (ChatGPT, Claude, Gemini) writin

Xue Bin Peng

MimicKit is an open-source framework for training motion controllers using motion imitation and reinforcement learning. The codebase provides implementations of commonly-used motion-imitation techniques and RL algorithms. This framework is intended to support research and applications in computer graphics and robotics by providing a unified training framewor

Nir Goren, Oren Katzir, Abhinav Nakarmi, Eyal Ronen

With the rapid adoption of diffusion models for visual content generation, proving authorship and protecting copyright have become critical. This challenge is particularly important when model owners keep their models private and may be unwilling or unable to handle authorship issues, making third-party verification essential. A natural solution is to embed

Ziqing Lu, Lifeng Lai, Weiyu Xu

Reinforcement learning (RL) for the Markov Decision Process (MDP) has emerged in many security-related applications, such as autonomous driving, financial decisions, and drone/robot algorithms. In order to improve the robustness/defense of RL systems against adversaries, studying various adversarial attacks on RL systems is very important. Most previous work

Coleman Drake, Mark K. Meiselbach, Daniel Polsky

Enrollment in the Health Insurance Marketplaces created by the Affordable Care Act reached an all-time high of approximately 25 million Americans in 2025, roughly doubling since enhanced premium tax credit subsidies were made available in 2021. The scheduled expiration of enhanced subsidies in 2026 is estimated to leave over seven million Americans without h

Victor Olkhov

We present the unified market-based description of returns and variances of the trades with shares of a particular security, of the trades with shares of all securities in the market, and of the trades with the market portfolio. We consider the investor who doesn't trade the shares of his portfolio he collected at time t0 in the past. The investor observes t

Md. Joshem Uddin, Soham Changani, Baris Coskunuzer

Temporal graph classification plays a critical role in applications such as cybersecurity, brain connectivity analysis, social dynamics, and traffic monitoring. Despite its significance, this problem remains underexplored compared to temporal link prediction or node forecasting. Existing methods often rely on snapshot-based or recurrent architectures that ei

Jaume de Haro, Emilio Elizalde

We present a novel derivation of the spacetime metric generated by matter, without invoking Einstein's field equations. For static sources, the metric arises from a relativistic formulation of D'Alembert's principle, where the inertial force is treated as a real dynamical entity that exactly compensates gravity. This leads to a conformastatic metric whose ge

Adam J. Czarnecki, Andrzej Czarnecki, Raquel Secrist, Julia Willsey

Videos of the 2020 Beirut explosion offer a rare opportunity to see a shock wave. We summarize the non-linear theory of a weak shock, derive the Landau-Whitham formula for the thickness of the overpressure layer and, using frame-by-frame video analysis, we demonstrate agreement of data and theory.

V. De Henau, S. Bouma, J. Bray, S. Buitink

Double-bump showers are a rare class of extensive air showers (EAS) predicted by Monte Carlo simulations. They occur when a high-energy secondary particle, the leading particle, travels significantly farther than the rest, creating a distinct double-peaked longitudinal profile. So far, no experiment has been able to directly detect these showers. The unique

Seyed Mohammad Mousavi, Morteza Analoui

Story continuation focuses on generating the next image in a narrative sequence so that it remains coherent with both the ongoing text description and the previously observed images. A central challenge in this setting lies in utilizing prior visual context effectively, while ensuring semantic alignment with the current textual input. In this work, we introd

Devvrit Khatri, Lovish Madaan, Rishabh Tiwari, Rachit Bansal

Reinforcement learning (RL) has become central to training large language models (LLMs), yet the field lacks predictive scaling methodologies comparable to those established for pre-training. Despite rapidly rising compute budgets, there is no principled understanding of how to evaluate algorithmic improvements for scaling RL compute. We present the first la

Stanisław Drożdż, Robert Kluszczyński, Jarosław Kwapień, Marcin Wątorek

Multifractality in time series analysis characterizes the presence of multiple scaling exponents, indicating heterogeneous temporal structures and complex dynamical behaviors beyond simple monofractal models. In the context of digital currency markets, multifractal properties arise due to the interplay of long-range temporal correlations and heavy-tailed dis

Ziyang Xiong, Tong Lin, Liu Li, Hao Deng

Integrated photonic circuits are foundational for versatile applications, where high-performance traveling-wave optical resonators are critical. Conventional whispering-gallery mode microresonators (WGMRs) confine light in closed-loop waveguide paths, thus inevitably occupy large footprints. Here, we report an ultracompact high loaded Q silicon photonic WGMR

Maciej T. Jarema, Mohammadamin Tajik, Jörg Schmiedmayer, Silke Weinfurtner

Information shared between parties quantifies their correlation. The encoding of correlations across space and time characterises the structure, history, and interactions of systems. One of the most fundamental properties that emerges from studies of information is the area law, which states that information shared between spatial subregions typically scales

B. D. MacNeil, J. S. R. McCoombs, D. Kalliecharan, J. Myra

We report a method to grow B20 MnGe thin films using molecular-beam epitaxy, which employs an ultrathin CrSi template layer on Si(111). This layer is expected to be nonmagnetic, in contrast to MnSi and FeGe buffer layers that have been used previously. This template layer permits an investigation of the intrinsic properties of MnGe in the ultrathin-film limi

Benjamin Z. Gregory, Neha Wadehra, Shuyuan Zhang, Yi Wu

Excitement about the magnetic and electronic properties of RuO$_2$ is growing, fueled by reports of antiferromagnetism, strain-induced superconductivity, and its recent classification as a member of a newly proposed magnetic class, altermagnets, with RuO$_2$ widely regarded as the paradigmatic example. Nevertheless, the magnetic ground state of RuO$_2$ remai

Yingzhi Tao, Chang Yang

This paper quantifies the age-stratified global burden of four mental disorders in 27 regions from 1990 to 2021 using GBD 2021. To put it in detail, it links the age-standardized years of disability adjustment with 18 world development indicators across economic, educational, social and information technology sectors. Then, by means of Pearson correlation, m

Xinyi Chen, Yilun Chen, Yanwei Fu, Ning Gao

We introduce InternVLA-M1, a unified framework for spatial grounding and robot control that advances instruction-following robots toward scalable, general-purpose intelligence. Its core idea is spatially guided vision-language-action training, where spatial grounding serves as the critical link between instructions and robot actions. InternVLA-M1 employs a t

Joshua Brakensiek, Yeyuan Chen, Manik Dhar, Zihan Zhang

In coding theory, a common question is to understand the threshold rates of various local properties of codes, such as their list decodability and list recoverability. A recent work Levi, Mosheiff, and Shagrithaya (FOCS 2025) gave a novel unified framework for calculating the threshold rates of local properties for random linear and random Reed--Solomon code

Benjamin de Jonge, Haocheng Zhang, Manel Errando, Andrea Gokus

The polarization of X-ray synchrotron emission in blazars directly probes the magnetic field geometry and particle acceleration processes in relativistic jets. We use particle-in-cell simulations of magnetic reconnection and magnetized turbulence, coupled to polarization-sensitive radiative transfer code, to interpret IXPE observations of Mrk 421 during a hi

Joshua Brakensiek, Yeyuan Chen, Manik Dhar, Zihan Zhang

In coding theory, the problem of list recovery asks one to find all codewords $c$ of a given code $C$ which such that at least $1-\rho$ fraction of the symbols of $c$ lie in some predetermined set of $\ell$ symbols for each coordinate of the code. A key question is bounding the maximum possible list size $L$ of such codewords for the given code $C$. In this

Iye Szin Ang, Martin Johannes Findl, Elisabeth Hauzinger, Klaus Philipp Sedlazeck

Automated rock classification from mineral composition presents a significant challenge in geological applications, with critical implications for material recycling, resource management, and industrial processing. While existing methods using One dimensional Convolutional Neural Network (1D-CNN) excel at mineral identification through Raman spectroscopy, th

Nicolas Billerey, Imin Chen, Lassina Dembélé, Luis Dieulefait

We solve the Fermat-type equation \[ x^{13} + y^{13} = 3 z^7, \qquad \gcd(x,y,z) = 1 \] combining a unit sieve, the multi-Frey modular method, level raising, computations of systems of eigenvalues modulo 7 over a totally real field, and results for reducibility of certain Galois representations.

Qiwei Yuan, Zhitong Xu, Yinghao Chen, Yiming Xu

Machine learning solvers for partial differential equations (PDEs) have attracted growing interest. However, most existing approaches, such as neural network solvers, rely on stochastic training, which is inefficient and typically requires a great many training epochs. Gaussian process (GP)/kernel-based solvers, while mathematical principled, suffer from sca

P. C. Barry, A. Prokudin, T. Anderson, C. Cocuzza

We present the first simultaneous global QCD analysis of unpolarized transverse momentum dependent (TMD) and collinear parton distribution functions (PDFs) in the proton. Our study incorporates data from deep-inelastic scattering, Drell-Yan, inclusive weak boson, $W$+\,charm, and jet production involving PDFs, as well as TMD Drell-Yan and $Z$-boson productio

Fengbin Zhu, Xiang Yao Ng, Ziyang Liu, Chang Liu

Deep Research (DR) agents, powered by advanced Large Language Models (LLMs), have recently garnered increasing attention for their capability in conducting complex research tasks. However, existing literature lacks a rigorous and systematic evaluation of DR Agent's capabilities in critical research analysis. To address this gap, we first propose HisRubric, a

Wan-Zhe Feng, Jinzheng Li, Pran Nath, Zong-Huan Ye

An analysis of baryogenesis and stochastic gravitational wave production is presented for an extension of the standard model where the dark sector consists of dark matter particles charged under a $U(1)_x$ gauge symmetry, while a subset of dark fields also carry lepton number but no $U(1)_x$ charge. We demonstrate that with CP violation induced by Yukawa cou

J. Olivo, J. Blengino Albrieu, Mauro Cuevas

We propose a model based on density functional theory (DFT) and quantum electrodynamics (QED) to study the dynamical characteristics of graphene quantum dots (GQDs). We assume the GQD edges are saturated with hydrogen atoms, effectively making it a polycyclic aromatic hydrocarbon (PAH) such as coronene. By combining the GQD spectrum calculated from a time-de

Yibo Peng, James Song, Lei Li, Xinyu Yang

Code agents are increasingly trusted to autonomously fix bugs on platforms such as GitHub, yet their security evaluation focuses almost exclusively on functional correctness. In this paper, we reveal a novel type of threat to real-world code agents: Functionally Correct yet Vulnerable (FCV) patches, which pass all test cases but contain vulnerable code. With

Connor Lane, Mihir Tripathy, Leema Krishna Murali, Ratna Sagari Grandhi

We study the problem of training self-supervised foundation models for functional MRI. Our main contributions are: (1) we introduce a new model family (CortexMAE) trained using the masked autoencoder framework on 2.1K hours of open fMRI data, and (2) we release the first open evaluation suite (Brainmarks) for fMRI foundation models. Our core innovation is si

Connor A Occhialini, Christie Nelson, Alessandro Bombardi, Shiyu Fan

We report Ru L$_3$-edge resonant X-ray diffraction studies on single crystal and (001) epitaxial films of RuO$_2$. We investigate the distinct $\mathbf{Q} = (100)$ and $(001)$ reflections as a function of incident energy, azimuthal angle, and temperature. The results show that the observed resonant diffraction in RuO$_2$ is fully consistent with a resonant c

Xena Al-Hejji, Santina Duarte, Jose Guillermo Gomez Castro, Edgar Bermudez Contreras

Deep reinforcement learning (DRL) algorithms have the potential to provide new insights into psychiatric disorders. Here we create a DRL model of schizophrenia: a complex psychotic disorder characterized by anhedonia, avoidance, temporal discounting, catatonia, and hallucinations. Schizophrenia's causes are not well understood: dopaminergic theories emphasiz

Anurag Garg

Photometric classification of Type Ia supernovae (SNe Ia) is critical for cosmological studies but remains difficult due to class imbalance and observational noise. While deep learning models have been explored, they are often resource-intensive and lack interpretability. We present a computationally efficient and interpretable classification framework that

Joshua Wang

The Rickard complex of a braid with strands colored by positive integers is a chain complex of singular Soergel bimodules. The complex determines the colored triply-graded homology and colored sl(N) homology of the braid closure, when closure is color-compatible. For each braid on two strands with any colors, we construct a minimal complex that is homotopy e

Yang Yang, Severi Rissanen, Paul E. Chang, Nasrulloh Loka

Amortized simulator-based inference offers a powerful framework for tackling Bayesian inference in computational fields such as engineering or neuroscience, increasingly leveraging modern generative methods like diffusion models to map observed data to model parameters or future predictions. These approaches yield posterior or posterior-predictive samples fo

Paolo Conti, Mengwu Guo, Attilio Frangi, Andrea Manzoni

Highly accurate datasets from numerical or physical experiments are often expensive and time-consuming to acquire, posing a significant challenge for applications that require precise evaluations, potentially across multiple scenarios and in real-time. Even building sufficiently accurate surrogate models can be extremely challenging with limited high-fidelit

Jonathan Nemirovsky, Lee Peleg, Amit Ben Kish, Yotam Shapira

We investigate quantum circuits built from arbitrary single-qubit operations combined with programmable all-to-all multiqubit entangling gates that are native to, among other systems, trapped-ion quantum computing platforms. We report a constant-cost of no more than 6 application of such Clifford entangling multiqubit gates to realize any sequence of Cliffor

Mikolaj Walczak, Uttej Kallakuri, Edward Humes, Xiaomin Lin

Vision Transformers (ViTs) have demonstrated strong capabilities in interpreting complex medical imaging data. However, their significant computational and memory demands pose challenges for deployment in real-time, resource-constrained mobile and wearable devices used in clinical environments. We introduce, BiTMedViT, a new class of Edge ViTs serving as med

Kai Zou, Ziqi Huang, Yuhao Dong, Shulin Tian

Unified multimodal models aim to jointly enable visual understanding and generation, yet current benchmarks rarely examine their true integration. Existing evaluations either treat the two abilities in isolation or overlook tasks that inherently couple them. To address this gap, we present Uni-MMMU, a comprehensive and discipline-aware benchmark that systema

Myoungkyu Lee, Yongyun Hwang

Statistical structure and the underlying energy budget of wall shear stress fluctuations are studied in both Poiseulle and Couette flows with emphasis on its streamwise component. Using a dimensional analysis and direct numerical simulation data, it is shown that the spectra of streamwise wall dissipation for $\lambda \lesssim 1000 \delta_\nu$ are asymptotic

Balázs Mészáros, James C. Knight, Jonathan Timcheck, Thomas Nowotny

Spiking Neural Networks are attracting increased attention as a more energy-efficient alternative to traditional Artificial Neural Networks for edge computing. Neuromorphic computing can significantly reduce energy requirements. Here, we present a complete pipeline: efficient event-based training of SNNs with synaptic delays on GPUs and deployment on Intel's

Junhong Shen, Mu Cai, Bo Hu, Ameet Talwalkar

Multimodal Large Language Models (MLLMs) struggle with precise reasoning for structured visuals like charts and diagrams, as pixel-based perception lacks a mechanism for verification. To address this, we propose to leverage derendering -- the process of reverse-engineering visuals into executable code -- as a new modality for verifiable visual reasoning. Spe

Timothé Albouy, Antonio Fernández Anta, Chryssis Georgiou, Nicolas Nicolaou

This paper explores necessary and sufficient system conditions to solve distributed tasks with binary outputs (\textit{i.e.}, tasks with output values in $\{0,1\}$). We focus on the distinct output sets of values a task can produce (intentionally disregarding validity and value multiplicity), considering that some processes may output no value. In a distribu

Manuel Mañas, Miguel Rojas

Uvarov-type perturbations for mixed-type multiple orthogonal polynomials on the step line are investigated within a matrix-analytic framework. The transformations considered involve both rational and additive modifications of a rectangular matrix of measures, implemented through left and right multiplication by regular matrix polynomials together with the ad

Lap Chi Lau, Akshay Ramachandran

A fundamental problem in statistics is estimating the shape matrix of an Elliptical distribution. This generalizes the familiar problem of Gaussian covariance estimation, for which the sample covariance achieves optimal estimation error. For Elliptical distributions, Tyler proposed a natural M-estimator and showed strong statistical properties in the asympto

Zhiqi Huang, Vivek Datla, Chenyang Zhu, Alfy Samuel

We propose a method for confidence estimation in retrieval-augmented generation (RAG) systems that aligns closely with the correctness of large language model (LLM) outputs. Confidence estimation is especially critical in high-stakes domains such as finance and healthcare, where the cost of an incorrect answer outweighs that of not answering the question. Ou

Ivan Vykopal, Matúš Pikuliak, Simon Ostermann, Marián Šimko

Chat assistants increasingly integrate web search functionality, enabling them to retrieve and cite external sources. While this promises more reliable answers, it also raises the risk of amplifying misinformation from low-credibility sources. In this paper, we introduce a novel methodology for evaluating assistants' web search behavior, focusing on source c

Thomas van Vuren, Fiona Sloothaak, Maarten G. Wolf, Jaron Sanders

The curse of dimensionality renders Reinforcement Learning (RL) impractical in many real-world settings with exponentially large state and action spaces. Yet, many environments exhibit exploitable structure that can accelerate learning. To formalize this idea, we study RL in Block Markov Decision Processes (BMDPs). BMDPs model problems with large observation

Wenwen Tong, Hewei Guo, Dongchuan Ran, Jiangnan Chen

We introduce InteractiveOmni, a unified and open-source omni-modal large language model for audio-visual multi-turn interaction, ranging from 4B to 8B parameters, designed to lead the field of lightweight models by offering comprehensive omni-modal understanding and speech generation capabilities. To achieve this, we integrate the vision encoder, audio encod

Surya Majumder, Daniel Widdowson, Yury Elkin, Olga Anosova

Ideal symmetry is known to break down under almost any noise. One measure of asymmetry in a periodic crystal is the relative multiplicity Z' of geometrically non-equivalent units. However, Z' discontinuously changes under almost any displacement of atoms, which can arbitrarily scale up a primitive cell. This discontinuity was recently resolved by a hierarchy

Tianshuo Xu, Kai Wang, Zhifei Chen, Leyi Wu

Computational replication of Chinese calligraphy remains challenging. Existing methods falter, either creating high-quality isolated characters while ignoring page-level aesthetics like ligatures and spacing, or attempting page synthesis at the expense of calligraphic correctness. We introduce \textbf{UniCalli}, a unified diffusion framework for column-level

Shrey Pandit, Austin Xu, Xuan-Phi Nguyen, Yifei Ming

Large language model (LLM)-based reasoning systems have recently achieved gold medal-level performance in the IMO 2025 competition, writing mathematical proofs where, to receive full credit, each step must be not only correct but also sufficiently supported. To train LLM-based reasoners in such challenging, open-ended settings, strong verifiers capable of ca

Efe Yazgan

Polarization and spin correlation measurements of top quark-antiquark ($t\bar{t}$) pairs provide tests of the standard model, but also new ways to test quantum mechanics with unstable particles at highest energies ever produced in a laboratory. Recent $t\bar{t}$ spin correlation measurements and the tests they enable, made with the CMS detector at the CERN L

Ganga R. Nair, V. Sreenath

Primordial magnetic fields (PMFs) are magnetic fields generated during the early universe. These fields are thought to be the seeds of extragalactic magnetic fields. The origin of PMFs is not well known. Further, if they are indeed sources of extragalactic fields, then there is a possibility that observations of extragalactic magnetic fields could provide in

Kiran Eiden, Daniel Kasen

A long-lived central engine embedded in expanding supernova ejecta can alter the dynamics and observational signatures of the event, producing an unusually luminous, energetic, and/or rapidly-evolving transient. We use two-dimensional hydrodynamics simulations to study the effect of a central energy source, varying the amount, rate, and isotropy of the energ

Mustafa Munir, Alex Zhang, Radu Marculescu

Vision graph neural networks (ViG) have demonstrated promise in vision tasks as a competitive alternative to conventional convolutional neural nets (CNN) and transformers (ViTs); however, common graph construction methods, such as k-nearest neighbor (KNN), can be expensive on larger images. While methods such as Sparse Vision Graph Attention (SVGA) have show

R. M. Ludlam, J. M. Miller, E. M Cackett, J. A. Garcia

We present the first XRISM/Resolve observation of the persistently accreting neutron star (NS) low-mass X-ray binary Serpens X-1. The source was observed on October 17th, 2024, for approximately 350 ks of elapsed time, resulting in 171 ks of exposure. The source exhibited 22% variability with respect to the average count rate of 73.1 count/s during the obser

Jingyi Zhou, Cheng Chen, Kai Zuo, Manjie Xu

Large language models (LLMs) have recently demonstrated strong potential for sequential recommendation. However, current LLM-based approaches face critical limitations in modeling users' long-term and diverse interests. First, due to inference latency and feature fetching bandwidth constraints, existing methods typically truncate user behavior sequences to i

Sarah Thiele, Romain Teyssier

Galactic cosmic rays (CRs) play a crucial role in galaxy formation and evolution by altering gas dynamics and chemistry across multiple scales. Typical numerical simulations of CR transport assume a constant diffusion coefficient for the entire galaxy, despite both numerical and theoretical studies showing that it can change by orders of magnitude depending

Pavan Vynatheya, Taeho Ryu, Chen Wang, Alison Sills

A significant fraction of stars experience close interactions, including collisions resulting from gravitational encounters and mergers within close binary systems. These processes can produce more massive stars that may give rise to relatively rare objects such as blue stragglers. Distinguishing the outcomes of collisions and mergers is challenging yet esse

Zhenxuan Zhang, Peiyuan Jing, Zi Wang, Ula Briski

Synthesizing high-quality images from low-field MRI holds significant potential. Low-field MRI is cheaper, more accessible, and safer, but suffers from low resolution and poor signal-to-noise ratio. This synthesis process can reduce reliance on costly acquisitions and expand data availability. However, synthesizing high-field MRI still suffers from a clinica

Xiuyuan Chen, Tao Sun, Dexin Su, Ailing Yu

Current benchmarks for AI clinician systems, often based on multiple-choice exams or manual rubrics, fail to capture the depth, robustness, and safety required for real-world clinical practice. To address this, we introduce the GAPS framework, a multidimensional paradigm for evaluating Grounding (cognitive depth), Adequacy (answer completeness), Perturbation

Amine Asselah, Vittoria Silvestri, Lorenzo Taggi

Internal Diffusion Limited Aggregation is an interacting particle system that describes the growth of a random cluster governed by the boundary harmonic measure seen from an internal point. Our paper studies IDLA in $\mathbb{Z}^d$ driven by critical branching random walks. We prove that, unlike classical IDLA, this process exhibits a phase transition in the

Mohd Saif Ali Khan, Karthik RM, Samar Agnihotri

Pilot contamination remains a major bottleneck in realizing the full potential of distributed massive MIMO systems. We propose two dynamic and scalable pilot assignment schemes designed for practical deployment in such networks. First, we present a low-complexity centralized scheme that sequentially assigns pilots to user equipments (UEs) to minimize the glo

Areen Khalaila, Dylan Cashman

Tactile graphics are often adapted from visual chart designs, yet many of these encodings do not translate effectively to non-visual exploration. Blind and low-vision (BLV) people employ a variety of physical strategies such as measuring lengths with fingers or scanning for texture differences to interpret tactile charts. These observations suggest an opport

Haoting Zhen, Yifei He, Sampriti Saha, Mithilesh K. Parit

Two-dimensional (2D) dipolar atomic gases present unique opportunities for exploring novel quantum phases due to their anisotropic and long-range interactions. However, the behavior of strongly dipolar Bose gases in 2D remains unclear, especially when dipoles are tilted. Here, we demonstrate the creation and characterization of strongly dipolar 2D condensate

Aymeric Fleith, Julian Zirbel, Daniel Cremers, Niclas Zeller

We present LiFMCR, a novel dataset for the registration of multiple micro lens array (MLA)-based light field cameras. While existing light field datasets are limited to single-camera setups and typically lack external ground truth, LiFMCR provides synchronized image sequences from two high-resolution Raytrix R32 plenoptic cameras, together with high-precisio

Andrzej Niedzielski, Robert Jaros, Artur Paczuski, Monika Adamów

Radial velocity searches may lead to detection of exoplanets at large orbital separations only if long-enough time-series of data are available. Therefore publication of precise measurements collected in the past is very valuable even if not successfully completed with a definitive detection. Here we present 309 precise ($\sigma$RV$\approx$5-7 m s$^{-1}$) mu

Adrian Kummerländer, Shota Ito, Maximilian Schecher, Davide Dapelo

Accurately capturing the dynamic forces acting on rotors as well as their wake effects presents a significant challenge for computational fluid dynamics (CFD) due to high Reynolds numbers and a large range of spatio-temporal scales. The present work proposes a novel blade-resolved wall-modeled large eddy simulation (WMLES) approach based on the lattice Boltz

Celia Mengyue Li, Sophie Pull, Steven Ramsay

We introduce a new two-sided type system for verifying the correctness and incorrectness of functional programs with atoms and pattern matching. A key idea in the work is that types should range over sets of normal forms, rather than sets of values, and this allows us to define a complement operator on types that acts as a negation on typing formulas. We sho

Aditya Tanikanti, Benoit Côté, Yanfei Guo, Le Chen

We present the Federated Inference Resource Scheduling Toolkit (FIRST), a framework enabling Inference-as-a-Service across distributed High-Performance Computing (HPC) clusters. FIRST provides cloud-like access to diverse AI models, like Large Language Models (LLMs), on existing HPC infrastructure. Leveraging Globus Auth and Globus Compute, the system allows

Megumi Ishida, Hiroshi Ohki, Shohei Uemura

We propose a simple and unified framework that simultaneously explains the origins of light Dirac neutrino masses, asymmetric dark matter (ADM), and the baryon asymmetry of the Universe. The model is based on an extended $U(1)_X$ Froggatt-Nielsen--like mechanism, which naturally generates suppressed Yukawa couplings and realizes a Dirac seesaw for neutrino m

Carlo Saccardi, Maximilian Pierzyna, Haitz Sáez de Ocáriz Borde, Simone Monaco

Kilometer-scale weather data is crucial for real-world applications but remains computationally intensive to produce using traditional weather simulations. An emerging solution is to use deep learning models, which offer a faster alternative for climate downscaling. However, their reliability is still in question, as they are often evaluated using standard m

Run Luo, Xiaobo Xia, Lu Wang, Longze Chen

Next-generation multimodal foundation models capable of any-to-any cross-modal generation and multi-turn interaction will serve as core components of artificial general intelligence systems, playing a pivotal role in human-machine interaction. However, most existing multimodal models remain constrained by autoregressive architectures, whose inherent limitati

Changsheng Wang, Xin Chen, Sijia Liu, Ke Ding

Adapting pretrained large language models (LLMs) to code domains via supervised fine-tuning (FT) has been commonly used for code generation. However, we identify a previously underappreciated failure mode, the memorization barrier, where strong memorization of downstream code data in the base model could trap optimization and prevent the standard FT from eff