October 2025 arXiv papers — page 91

Jinyu Hu, Chen Wu

Recently, building upon the research findings of E. L. Medeiros, we have extended the alpha-particle non-locality effect to the two-potential approach (TPA). This extension demonstrates that the integration of the alpha-particle nonlocality effect into TPA yields relatively favorable results. In the present work, we employ machine learning methods to further

Haoran Chen, Ruxuan Liu, Gedalia Y. Koehler, Fatemehsadat Tabatabaei

Integrated photonics has undergone tremendous development in the past few decades, transforming many fields of study in science and technology. Loss and gain are two fundamental elements in photonic circuits and have direct impacts on nearly all key performance metrics. Surprisingly, the tools to characterize the optical loss and gain inside photonic integra

Florian Lehner, Benjamin Shirley

We study the problem of whether rectangular polyominoes with holes are cube-foldable, that is, whether they can be folded into a cube, if creases are only allowed along grid lines. It is known that holes of sufficient size guarantee that this is the case. Smaller holes which by themselves do not make a rectangular polyomino cube-foldable can sometimes be com

Yoshinari Fujinuma

Large Language Models (LLMs) are commonly used as evaluators in various applications, but the reliability of the outcomes remains a challenge. One such challenge is using LLMs-as-judges for direct assessment, i.e., assigning scores from a specified range without any references. Focusing on summarization, we first show that this challenge stems from LLM judge

Jostein Barry-Straume, Adwait D. Verulkar, Arash Sarshar, Andrey A. Popov

The objective of designing a control system is to steer a dynamical system with a control signal, guiding it to exhibit the desired behavior. The Hamilton-Jacobi-Bellman (HJB) partial differential equation offers a framework for optimal control system design. However, numerical solutions to this equation are computationally intensive, and analytical solution

Bo-Hae Im, Hansol Kim

Let $K$ be a field of characteristic $0$ and $E/K$ an elliptic curve over $K$. For a finite extension $L/K$ and a prime~$\ell$, we provide Galois-theoretic sufficient conditions on $L/K$ under which $E\left(L\right)\left[\ell^{\infty}\right] = E\left(K\right)\left[\ell^{\infty}\right]$. For a non-Galois extension $L/K$ of prime degree, we relate the growth o

Keivan Shariatmadar, Ahmad Osman, Ramin Ray, Kisam Kim

Fair, transparent, and explainable decision-making remains a critical challenge in Olympic and Paralympic combat sports. This paper presents \emph{FST.ai 2.0}, an explainable AI ecosystem designed to support referees, coaches, and athletes in real time during Taekwondo competitions and training. The system integrates {pose-based action recognition} using gra

Hadis Rezaei, Ahmed Afif Monrat, Karl Andersson, Francesco Flammini

The inherent determinism of blockchain technology poses a significant challenge to generating secure random numbers within smart contracts, leading to exploitable vulnerabilities, particularly in decentralized finance (DeFi) ecosystems and blockchain-based gaming applications. From our observations, the current state-of-the-art detection tools suffer from in

Isaac Viviano

A procedure is presented to estimate the diffusion coefficient of a uniform patch of argon gas in a uniform background of helium gas. Molecular Dynamics (MD) simulations of the two gases interacting through the Lennard-Jones potential are carried out using the LAMMPS software package. In addition, finite-difference (FD) calculations are used to solve the con

Zhanhong He, Hanyu Meng, David Huang, Roberto Togneri

Estimating piano dynamic from audio recordings is a fundamental challenge in computational music analysis. In this paper, we propose an efficient multi-task network that jointly predicts dynamic levels, change points, beats, and downbeats from a shared latent representation. These four targets form the metrical structure of dynamics in the music score. Inspi

Chengrun Li, Corentin Royer, Haozhe Luo, Bastian Wittmann

Most current medical vision language models struggle to jointly generate diagnostic text and pixel-level segmentation masks in response to complex visual questions. This represents a major limitation towards clinical application, as assistive systems that fail to provide both modalities simultaneously offer limited value to medical practitioners. To alleviat

Fatima AlGhamdi, Omar Alharbi, Abdullah Aldwyish, Raied Aljadaany

Detecting anomalies in crowded scenes is challenging due to severe inter-person occlusions and highly dynamic, context-dependent motion patterns. Existing approaches often struggle to adapt to varying crowd densities and lack interpretable anomaly indicators. To address these limitations, we introduce VelocityNet, a dual-pipeline framework that combines head

Alexander Kolpakov

We introduce a frequency-tunable, two-dimensional non-Abelian control of operation order constructed from the reduced Burau representation of the braid group $B_3$, specialised at $t=e^{i\omega}$ and unitarized by Squier's Hermitian form. Coupled to two non-commuting qubit unitaries $A$, $B$, the resulting switch admits a closed expression for the single-sho

Karelia Salinas, Luis Gustavo Nonato, Jean-Daniel Fekete, Fernanda Bartolo dos Santos Saran

We propose two novel interaction techniques for visualization-assisted exploration of urban data: Layer Toggling and Visibility-Preserving Lenses. Layer Toggling mitigates visual overload by organizing information into separate layers while enabling comparisons through controlled overlays. This technique supports focused analysis without losing spatial conte

Lukas Helff, Ruben Härle, Wolfgang Stammer, Felix Friedrich

Large language models (LLMs) excel at generating fluent text, but their internal reasoning remains opaque and difficult to control. Sparse autoencoders (SAEs) make hidden activations more interpretable by exposing latent features that often align with human concepts. Yet, these features are fragile and passive, offering no mechanism for systematic reasoning

Shriyansh Agrawal, Aidan Lau, Sanyam Shah, Ahan M R

The prevalence of Large Language Models (LLMs) for generating multilingual text and source code has only increased the imperative for machine-generated content detectors to be accurate and efficient across domains. Current detectors, predominantly utilizing zero-shot methods, such as Fast DetectGPT or GPTZero, either incur high computational cost or lack suf

Ji-Hoon Ha, Elena S. Volnova

Density inhomogeneities are ubiquitous in space and astrophysical plasmas, particularly at magnetic reconnection sites, shock fronts, and within compressible turbulence. The gradients associated with these inhomogeneous plasma regions serve as free energy sources that can drive plasma instabilities, including the lower-hybrid drift instability (LHDI). Notabl

Sin Kyu Kang, C. J. Ouseph

The neutrino dipole portal (NDP) is a minimal and predictive extension of the Standard Model, in which a transition magnetic moment operator couples an active neutrino to a heavy neutral lepton via the electromagnetic field. This higher-dimensional interaction gives rise to distinctive processes such as neutrino up-scattering, radiative decays, meson transit

Vincent Cohen-Addad, David P. Woodruff, Shenghao Xie, Samson Zhou

We study the problem of graph and hypergraph sparsification in insertion-only data streams. The input is a hypergraph $H=(V, E, w)$ with $n$ nodes, $m$ hyperedges, and rank $r$, and the goal is to compute a hypergraph $\widehat{H}$ that preserves the energy of each vector $x \in \mathbb{R}^n$ in $H$, up to a small multiplicative error. In this paper, we give

Rui Jerry Huang, Wendy Liu, Anastasia Miin, Lei Ding

Inference-time computation is a critical yet challenging paradigm for enhancing the reasoning performance of large language models (LLMs). While existing strategies improve reasoning stability and consistency, they suffer from notable limitations: self-correction often reinforces the model's initial biases, and Multi-Agent Collaboration (MAC) often fails due

Jack Yang, Ziqi Yin, Lei Ao, Sean Li

Recent developments in materials informatics and artificial intelligence has led to the emergence of foundational energy models for material chemistry, as represented by the suite of MACE-based foundation models, bringing a significant breakthrough in universal potentials for inorganic solids. As to all method developments in computational materials science,

Sepehr Assadi, Janani Sundaresan, Helia Yazdanyar

We study graph coloring problems in the streaming model, where the goal is to process an $n$-vertex graph whose edges arrive in a stream, using a limited space that is smaller than the trivial $O(n^2)$ bound. While prior work has largely focused on coloring graphs with a large number of colors, we explore the opposite end of the spectrum: deciding whether th

Pavel Orlov, Cheryne Jonay, Tomaž Prosen

In this work, we introduce a broad class of circuits, or quantum cellular automata, which we call 'pairwise-difference-conserving circuits' (PDC). These models are characterized by local gates that preserve the pairwise difference of local operators (e.g. particle number). Such circuits can be de- fined on arbitrary graphs in arbitrary dimensions for

Harikrishna Sahu, Wei Xiong, Anagha Savit, Shivank S Shukla

Traditional machine learning has advanced polymer discovery, yet direct generation of chemically valid and synthesizable polymers without exhaustive enumeration remains a challenge. Here we present polyT5, an encoder-decoder chemical language model based on the T5 architecture, trained to understand and generate polymer structures. polyT5 enables both proper

Barak Kol

This article offers a broad-brush account of the Newtonian three-body problem, from its origins with Newton to its vibrant present, emphasizing its enduring influence on theoretical physics. It unfolds through a series of self-contained episodes that illuminate the scientific fields and the paradigm shift that have grown out of this problem.

Daniel Rodriguez-Guillen, Carlos Wiechers, Lorena Velazquez-Ibarra

We introduce a topological photonic-crystal fiber that embeds a short Su-Schrieffer-Heeger (SSH) chain and supports two edge supermodes. Using full-vector modal analysis and a coupled generalized nonlinear Schroedinger equation, we show that each supermode provides an independent nonlinear channel with a distinct broadening mechanism: the even supermode feat

Ipsita Mandal

Fermi arcs represent the surface states at the boundary of a three-dimensional topological semimetal with the vacuum, illustrating the notion of bulk-boundary correspondence playing out in real materials. Their special character is tied up with the topological charges carried by the nodes of the semimetal in the momentum space, where two or more bands cross.

Maximilian Hofer, Christopher Fuchs, Moritz Siebert, Christian Berger

Even small electrostatic potentials can dramatically influence the band structure of narrow-, broken-, and inverted-gap materials. A quantitative understanding often necessitates a self-consistent Hartree approach. The valence and conduction band states strongly hybridize and/or cross in these systems. This makes distinguishing between electrons and holes im

Kyo Kuroki, Yasuyuki Okoshi, Thiem Van Chu, Kazushi Kawamura

This paper proposes a novel matrix quantization method, Binary Quadratic Quantization (BQQ). In contrast to conventional first-order quantization approaches, such as uniform quantization and binary coding quantization, that approximate real-valued matrices via linear combinations of binary bases, BQQ leverages the expressive power of binary quadratic express

Luigi Quarantiello, Elia Piccoli, Jack Bell, Malio Li

The birth of Foundation Models brought unprecedented results in a wide range of tasks, from language to vision, to robotic control. These models are able to process huge quantities of data, and can extract and develop rich representations, which can be employed across different domains and modalities. However, they still have issues in adapting to dynamic, r

Muhammad Salman Wahidi, Maurice Pfeiffer, Xinyan Wu, Fatemeh Ebrahimi

Nanoporous gold (npAu) has emerged as a potential candidate for many optical applications exploiting its large surface to volume ratio and high broadband absorption. However, the physical origin of its enhanced visible and near infrared absorption remained unclear and till now was not explicable by simplified models. Here, we have employed leveled-wave appro

Marc Miranda-Riaza, Pierpaolo Fontana, Alessio Celi

The classical and quantum simulation of lattice gauge theories (LGTs) with Lie groups is hindered by the infinite-dimensional Hilbert space of gauge degrees of freedom. In a recent work [Phys. Rev. X 15, 031065 (2025)], we introduced a new truncation scheme -- here renamed as Renormalized Dual Basis (RDB) -- based on the resolution of the single-plaquette pr

Ma-Ke Yuan, Mingyi Li, Yang Zhou

We study the multipartite entanglement structure of quantum states prepared by the Euclidean path integral over three-manifolds with multiple torus boundaries (the so-called link states) in both Abelian and non-Abelian Chern-Simons theories. For three-component link states in the Abelian theory, we derive an explicit formula for the Rényi multi-entropy in te

David Eppstein

The subdivided double construction on 4-regular graphs was used by Potočnik and Wilson to explore semi-symmetric (edge-transitive but not vertex-transitive) graphs, and can be used to construct every semi-symmetric 4-regular graph that contains a pair of twin vertices. We show that (regardless of symmetry) subdivided doubles have another curious property: th

Haruki Yagi, Zongping Gong

A rigorous analysis is presented for the entanglement spectrum of quantum many-body states possessing a higher-form group-representation symmetry generated by topological Wilson loops, which is generally non-invertible. A general framework based on elementary algebraic topology and category theory is developed to determine the block structure of reduced dens

Zhi-Chong Qiu, Ye-Huang Pang, Qing-Guo Huang

We explored a parameterized slow-roll inflationary model within the $Λ$CDM framework, utilizing a combination of data from Planck 2018, ACT DR6, DESI DR2, and BICEP/Keck 2018 (P-ACT-LB-BK18). Additionally, we incorporated the SH0ES prior on $H_0$ (P-ACT-LB-BK18-$H_0$) to analyze the model within the early dark energy (EDE) framework. While the model with a p

Jonathan A. Rebouças, Celio R. Muniz, Francisco Bento Lustosa, Edson Otoniel

This work presents and investigates novel traversable wormhole solutions within the framework of Asymptotically Safe Gravity (ASG), sourced by a dark matter halo modeled by the Dekel--Zhao density profile. The scale-dependent gravitational coupling $G(k)$, derived from the ASG renormalization group flow in the infrared regime, is incorporated directly into t

Julien Zouein, Hossein Javidnia, François Pitié, Anil Kokaram

We repurpose AV1 motion vectors to produce dense sub-pixel correspondences and short tracks filtered by cosine consistency. On short videos, this compressed-domain front end runs comparably to sequential SIFT while using far less CPU, and yields denser matches with competitive pairwise geometry. As a small SfM demo on a 117-frame clip, MV matches register al

Yanqiu Zou, Anastasiia Sapunova, Tommaso Giovannini, Chen Wang

Nanoporous metals are extensively investigated as platforms for applications in plasmonics. They present high surface areas and strong local electric fields that can be tuned at different energies, playing with the choice of the metals and the morphology of the porous layers. Until recently, research in the field of plasmonics has primarily focused on porous

Yandong Guan, Xilin Wang, Ximing Xing, Jing Zhang

In this work, we introduce CAD-Coder, a novel framework that reformulates text-to-CAD as the generation of CadQuery scripts - a Python-based, parametric CAD language. This representation enables direct geometric validation, a richer modeling vocabulary, and seamless integration with existing LLMs. To further enhance code validity and geometric fidelity, we p

Hisato Matsukawa

We introduce the relative Matsui spectrum, a new invariant associated with a stable \(\infty\)-category equipped with an action. This construction generalizes both Balmer's tensor triangular spectra and Matsui's triangular spectra, and provides a unified framework for classifying thick submodules. We establish its fundamental properties, including un

Kazimierz Chomicz

Let $K$ be a number field and $\mathfrak{p} \mid (2)$ be a prime ideal. We compute the fourth level of the $\mathfrak{p}$-adic completions of $K$ when the ramification index is $4$ and the inertial degree is trivial for the ideal $\mathfrak{p}$. This enables the computation of the fourth level of any $\mathfrak{p}$-adic completion of any quartic number field

Liangyang Liu, Yidian Li, Hengxin Tan, Yi Liu

While correlated phenomena of flat bands have been extensively studied in twisted systems, the ordered states that emerge from interactions in the intrinsic flat bands of kagome lattice materials remain largely unexplored. The newly discovered kagome metal CsCr3Sb5 offers a unique and rich platform for this research, as its multi-orbital flat bands at the Fe

Klaus Dohmen, Mandy Lange-Geisler

We present four combinatorial proofs of Morgado's formula for the number $\varrho(n)$ of non-congruent regular integers modulo $n$, corresponding to sequence A055653 in the On-Line Encyclopedia of Integer Sequences (OEIS), where an integer $m$ is said to be regular modulo $n$ if the congruence $m^2 x \equiv m \pmod{n}$ has a solution $x\in\mathbb{Z}$. To

Soumya Rani Samineni, Durgesh Kalwar, Vardaan Gangal, Siddhant Bhambri

Reinforcement Learning with Verifiable Rewards (RLVR)-based post-training of Large Language Models (LLMs) has been shown to improve accuracy on reasoning tasks and continues to attract significant attention. Existing RLVR methods, however, typically treat all tokens uniformly without accounting for token-level advantages. These methods primarily evaluate per

Pedro Cisneros-Velarde

We consider the situation where we have a limited number of denoising steps, i.e., of evaluations of a diffusion model. We show that two parallel processors or samplers under such limitation can improve the quality of the sampled image. Particularly, the two samplers make denoising steps at successive times, and their information is appropriately integrated

Kevin Coulembier, Serina Hu

We study polynomial functors in the incompressible category $\text{Ver}_4^+$, which can be viewed as super polynomial functors in characteristic 2. Concretely, we classify additive, exact and simple polynomial functors, and describe how simple polynomial functors evaluate on arbitrary objects. We also determine which objects are not annihilated by any polyno

Roman Malyshev, Ingeborg-Helene Svenum, Sverre M. Selbach, Thomas Tybell

A first principles study is conducted to explore the changes in the magnetic structure of Mn\textsubscript{3}GaN under anisotropic biaxial strain. Mn\textsubscript{3}GaN is an antiperovskite with a structure similar to that of an ideal cubic perovskite. Several manganese nitride antiperovskites including Mn\textsubscript{3}GaN were reported to have a frustra

Ritam Upadhyay, Naman Ahuja, Rishabh Baral, Aparna Garimella

Large language model (LLM) approaches to tabular summarization rely on extensive prompt engineering, decomposition pipelines, or entity-level intermediate representations to achieve strong performance. While effective, these strategies are computationally expensive and offer limited insight into how well models maintain state over long, evolving narratives.

Clementine Grethen, Simone Gasparini, Geraldine Morin, Jeremy Lebreton

Accurate 3D reconstruction of lunar surfaces is essential for space exploration. However, existing stereo vision reconstruction methods struggle in this context due to the Moon's lack of texture, difficult lighting variations, and atypical orbital trajectories. State-of-the-art deep learning models, trained on human-scale datasets, have rarely been tested on

Dario Olmos Cayo, Zui Oporto Almaraz, M. L. Peñafiel

We study the dynamical and thermodynamical stability of thin shells in (2+1)-dimensional spacetimes composed of an inner anti-de Sitter (AdS) region and an outer region described by a charged Ba\~nados--Teitelboim--Zanelli (BTZ) spacetime, sourced either by Einstein--Maxwell theory (Maxwell-BTZ) or Einstein--Born--Infeld theory (BI-BTZ). Assuming a fixed cha

Manik Rana, Calissa Man, Anotida Expected Msiiwa, Jeffrey Paine

Goal changes are a defining feature of real world multi-turn interactions, yet current agent benchmarks primarily evaluate static objectives or one-shot tool use. We introduce AgentChangeBench, a benchmark explicitly designed to measure how tool augmented language model agents adapt to mid dialogue goal shifts across three enterprise domains. Our framework f

Yu-Wen Chen, William Ho, Sasha M. Vergez, Grace Flaherty

The growing demand for home healthcare calls for tools that can support care delivery. In this study, we explore automatic health assessment from voice using real-world home care visit data, leveraging the diverse patient information it contains. First, we utilize Large Language Models (LLMs) to integrate Subjective, Objective, Assessment, and Plan (SOAP) no

Tomoyuki Ikeda, Shuji Machihara, Hayato Miyazaki, Tohru Ozawa

We revisit the derivation of the virial identity for nonlinear Schr\"odinger equations. In \cite{O06, FM17}, several conservation laws, such as for the charge and the energy, were derived without constructing a sequence of approximate solutions. Their approach involves additional properties of solutions due to Strichartz' estimate. In this paper, we derive t

Robert Griffiths

We consider a class of Gaussian Free Fields denoted by $(g_x)_{x \in {\cal V}_N}$, where $ {\cal V}_N = \{0,1\}^N$ and $N\in \mathbb{Z}_+$. These fields are related to a general class of $N$-dimensional random walks on the hypercube, which are killed at a certain rate. The covariance structure of the Gaussian free field is determined by the Green function of

Yining Xuan, Daito Miyazaki, Yuki Ishikawa, Mark Sadgrove

We demonstrate that non-chiral nanoparticles can produce chiral light when point emitters are coupled to their surface plasmon modes (SPMs) under certain conditions. Chiral emission arises from asymmetrical plasmon mode propagation from the source combined with the spin-momentum locked nature of the SPMs. The Purcell regime of cavity quantum electrodynamics

Yihong Dong, Zhaoyu Ma, Xue Jiang, Zhiyuan Fan

Diffusion language models (DLMs) are emerging as a compelling alternative to the dominant autoregressive paradigm, offering inherent advantages in parallel generation and bidirectional context modeling. However, for the tasks with strict structural constraints such as code generation, DLMs face a critical trade-off between inference speed and output quality,

Harry Buhrman, Sevag Gharibian, Zeph Landau, François Le Gall

We present an extremely simple polynomial-space exponential-time $(1-\varepsilon)$-approximation algorithm for MAX-k-SAT that is (slightly) faster than the previous known polynomial-space $(1-\varepsilon)$-approximation algorithms by Hirsch (Discrete Applied Mathematics, 2003) and Escoffier, Paschos and Tourniaire (Theoretical Computer Science, 2014). Our al

Emily Heath, Joseph Hyde, Natasha Morrison, Shannon Ogden

Let $k \ge 2$ and let $\bf G = \{G_1, \ldots, G_{m}\}$ be a collection of graphs on a common vertex set of cardinality $n$. We show that if each graph in $\bf G$ has minimum degree at least $(1-\frac{1}{2k} + o(1))n$, then for every edge-colouring $\chi$ of the $k$th power of a Hamilton cycle $C_n^k$ with $m$ colours, there is a copy of $C_n^k$ in $\bf G$ su

Yohei Ikenoue, Hitomi Tashiro, Shigeru Kuroyanagi

Prompt engineering is crucial for achieving reliable and effective outputs from large language models (LLMs), but its design requires specialized knowledge of prompting techniques and a deep understanding of target tasks. To address this challenge, we propose a novel method that adaptively selects task-appropriate prompting techniques based on users' abstrac

Hamsa Bastani, Osbert Bastani, Bryce McLaughlin

There has been a surge of recent interest in automatically learning policies to target treatment decisions based on rich individual covariates. In addition, practitioners want confidence that the learned policy has better performance than the incumbent policy according to downstream policy evaluation. However, due to the winner's curse -- an issue where the

Pranshav Gajjar, Molham Khoja, Abiodun Ganiyu, Marc Juarez

The impending adoption of Open Radio Access Network (O-RAN) is fueling innovation in the RAN towards data-driven operation. Unlike traditional RAN where the RAN data and its usage is restricted within proprietary and monolithic RAN equipment, the O-RAN architecture opens up access to RAN data via RAN intelligent controllers (RICs), to third-party machine lea

Andrey Itkin

This paper introduces a semi-analytical method for pricing American options on assets (stocks, ETFs) that pay discrete and/or continuous dividends. The problem is notoriously complex because discrete dividends create abrupt price drops and affect the optimal exercise timing, making traditional continuous-dividend models unsuitable. Our approach utilizes the

Yixue Cai, Xiyan Su, Dongpeng Yao, Rongduo Han

Mental health assessments are of central importance to individuals' well-being. Conventional assessment methodologies predominantly depend on clinical interviews and standardised self-report questionnaires. Nevertheless, the efficacy of these methodologies is frequently impeded by factors such as subjectivity, recall bias, and accessibility issues. Furthermo

Eric Keto, Abraham Loeb

A previous study suggests that the observed exponential character of the surface brightness profiles in the coma around the interstellar comet 3I/ATLAS at 4 au can be explained as a consequence of the destruction of the icy scattering particles by sublimation. Here we follow the evolution of the ice coma as a function of heliocentric distance. We describe th

Xiaoyan Xie, Anna Chen, Crisel Suarez, Katharine K. Reeves

We investigate the thermal properties of a solar flare by the observations of soft X-ray Telescope (XRT) on board Hinode and the Atmospheric Image Assembly on board the Solar Dynamics Observatory (SDO/AIA). Our analysis reveals a tenuous but hot plasma cloud located above the flare loops. The flare loops and plasma cloud exhibit independent temperature profi

Man-Lin Chu, Lucian Terhorst, Kadin Reed, Tom Ni

Simulating consumer decision-making is vital for designing and evaluating marketing strategies before costly real-world deployment. However, post-event analyses and rule-based agent-based models (ABMs) struggle to capture the complexity of human behavior and social interaction. We introduce an LLM-powered multi-agent simulation framework that models consumer

Antonio-Gabriel Chacón Menke, Phan Xuan Tan, Eiji Kamioka

Recent work has highlighted the importance of monitoring chain-of-thought reasoning for AI safety; however, current approaches that analyze textual reasoning steps can miss subtle harmful patterns and may be circumvented by models that hide unsafe reasoning. We present a sentence-level labeled dataset that enables activation-based monitoring of safety behavi

Ajay S. Gill, Yilun Guan, Adam D. Hincks, Tony Mroczkowski

We present a multifrequency and multi-instrument methodology to study the physical properties of galaxy clusters and cosmic filaments using cosmic microwave background observations. Our approach enables simultaneous measurement of both the thermal (tSZ) and kinematic Sunyaev-Zeldovich (kSZ) effects, incorporates relativistic corrections, and models astrophys

Zhuoyu Yao, Yue Wang, Songyang Zhang, Yingshu Li

Recent advances in distributed learning systems have introduced effective solutions for implementing collaborative artificial intelligence techniques in wireless communication networks. Federated learning approaches provide a model-aggregation mechanism among edge devices to achieve collaborative training, while ensuring data security, communication efficien

Ahmad M. Alkadri, Tyler D. Kharazi, K. Birgitta Whaley, Kranthi K. Mandadapu

The finite element method (FEM) is a cornerstone numerical technique for solving partial differential equations (PDEs). Here, we present $\textbf{Qu-FEM}$, a fault-tolerant era quantum algorithm for the finite element method. In contrast to other quantum PDE solvers, Qu-FEM preserves the geometric flexibility of FEM by introducing two new primitives, the uni

Wenlu Tang, Hongni Wang, Xingcai Zhou, Bei Jiang

We develop a novel approach to tackle the common but challenging problem of conformal inference for missing data in machine learning, focusing on Missing at Random (MAR) data. We propose a new procedure Conformal prediction for Missing data under Multiple Robust Learning (CM--MRL) that combines split conformal calibration with a multiple robust empirical-lik

Dan Friedman, Adithya Bhaskar, Alexander Wettig, Danqi Chen

Mechanistic interpretability strives to explain model behavior in terms of bottom-up primitives. The leading paradigm is to express hidden states as a sparse linear combination of basis vectors, called features. However, this only identifies which text sequences (exemplars) activate which features; the actual interpretation of features requires subjective in

William Lugoloobi, Chris Russell

Large language models exhibit a puzzling inconsistency: they solve complex problems yet frequently fail on seemingly simpler ones. We investigate whether LLMs internally encode problem difficulty in a way that aligns with human judgment, and whether this representation tracks generalization during reinforcement learning post-training. We train linear probes

Konrad Aguilar, Samantha Brooker, Jack Spielberg

The Effros-Shen algebra corresponding to an irrational number $\theta$ can be described by an inductive sequence of direct sums of matrix algebras, where the continued fraction expansion of $\theta$ encodes the dimensions of the summands, and how the matrix algebras at the $n$th level fit into the summands at the $(n+1)$th level. In recent work, Mitscher and

Sakineh Vosoughi-nia, Michał P. Nowak

Altermagnets provide a new route to engineer superconducting circuits without magnetic fields. We theoretically study the Andreev bound state (ABS) spectrum of a finite-width altrmagnet-based Josephson junction and show how the $d$-wave altermagnetic symmetry and geometric confinement shape its low-energy excitations. We find a clear distinction between the

Subin Mali, Yufei Zhao, Yu Wang, Saugata Sarker

X-ray and neutron diffraction are foundational tools for determining crystal structures, but their resolution limits can lead to misassignments, especially in materials with subtle distortions or competing phases. Here, we demonstrate the use of nonlinear transport as a complementary approach to uncover hidden crystal symmetries, using the strongly correlate

Huan Song, Deeksha Razdan, Yiyue Qian, Arijit Ghosh Chowdhury

Small Language Models (SLMs) offer compelling advantages in deployment cost and latency, but their accuracy often lags behind larger models, particularly for complex domain-specific tasks. While supervised fine-tuning can help bridge this performance gap, it requires substantial manual effort in data preparation and iterative optimization. We present PaDA-Ag

M. Grant Roberts, Wolfgang Altmannshofer, Pierce Giffin, Stefano Profumo

We develop a minimal, testable framework for two-component self-interacting dark matter (SIDM) in which a dominant, moderately self-interacting species coexists with an ultra-strongly self-interacting subcomponent (uSIDM). A light vector mediator induces velocity-dependent self-scattering, while early-universe dynamics - standard $2 \to 2$ annihilation suppl

Jiangyifei Zhu, Kuang Yuan, Akarsh Prabhakara, Yunzhi Li

Pulse Transit Time (PTT) is a measure of arterial stiffness and a physiological marker associated with cardiovascular function, with an inverse relationship to diastolic blood pressure (DBP). We present the first AI-enabled mmWave system for contactless multi-site PTT measurement using a single radar. By leveraging radar beamforming and deep learning algorit

David Schaich, Angel Sherletov

We present the latest results from our ongoing lattice field theory investigations of maximally supersymmetric Yang--Mills theory in three space-time dimensions, focusing on its non-perturbative phase diagram. Exploiting a lattice formulation that preserves a subset of the supersymmetry algebra at non-zero lattice spacing, we study the 'spatial deconfinement

Joe Bhangal, Allison W. S. Man, Tom J. L. C. Bakx, Darko Donevski

Studying the environments of dusty star-forming galaxies (DSFGs) provides insight into whether these luminous systems are reliable signposts of large-scale overdensities. Evidence suggests that individual DSFGs can trace overdense environments, although this association may not be universal. To test this, we investigate the environments surrounding two lumin

Arianna Tinari

Flavor-changing neutral current decays such as $b \to s \bar{\ell} \ell$ are highly suppressed in the Standard Model (SM) and therefore provide sensitive tests for new physics. Persistent tensions between SM predictions and experimental results in branching ratios and angular observables can be explained by a shift of the Wilson coefficient $C_9$ of the effe

Hrishikesh Sathyanarayan, Victor Vantilborgh, Ian Abraham

In this paper, we investigate the utility of datasets and whether more data or the 'right' data is advantageous for robot learning. In particular, we are interested on quantifying the utility of contact-based data as contact holds significant information for robot learning. Our approach derives a contact-aware objective function for learning object dynamics

Myles B. Sherman, Nikita Kosogorov, Casey Law, Vikram Ravi

We describe the design and commissioning tests for the DSA-110 Not-So-Fast Radio Burst (NSFRB) search pipeline, a 1.4GHz image-plane single-pulse search sensitive to 134ms-160.8s radio bursts. Extending the pulse width range of the FRB search by 3 orders of magnitude, the NSFRB search is sensitive to the recently-discovered Galactic Long Period Radio Transie

Jiahan Zhang, Muqing Jiang, Nanru Dai, Taiming Lu

Generative world models (WMs) can now simulate worlds with striking visual realism, which naturally raises the question of whether they can endow embodied agents with predictive perception for decision making. Progress on this question has been limited by fragmented evaluation: most existing benchmarks adopt open-loop protocols that emphasize visual quality

Soheil Abbasloo

What does it truly mean for a language model to "reason"? Current evaluations reward models' correct standalone answers-but correctness alone reveals little about the process that produced them. We argue that reasoning should be understood not as a static chain of steps but as a dynamic trajectory in which ideas interact, clash, and evolve into integrated in

Laura Thapa, Marybeth Arcodia, Elizabeth A. Barnes

We discuss the utility of applying clustering as a preprocessing step for identifying subseasonal to seasonal forecasts of opportunity of coastal sea level using convolutional neural networks (CNNs). Clustering leverages potential covariance among points along the same coastline or in the same ocean basin. To evaluate the utility of clustering for reliably i

Hugh Carvill

We study the Beurling Nyman (BN) family $f_\theta(x) = \{\theta/x\} - \theta\{1/x\}$ in $L^2((0,1])$ through a multiscale ladder parameterisation $\theta_{j,k} = 2^{-j}3^{-k}$ and the associated Gram matrix structure indexed by ladder distance. Using Mellin analysis and a controlled smoothing operator, we establish a rigorous polynomial decay envelope for of

Chengquan Guo, Yuzhou Nie, Chulin Xie, Zinan Lin

As large language models (LLMs) are increasingly used for code generation, concerns over the security risks have grown substantially. Early research has primarily focused on red teaming, which aims to uncover and evaluate vulnerabilities and risks of CodeGen models. However, progress on the blue teaming side remains limited, as developing defense requires ef

Jan Quan, Johan Suykens, Panagiotis Patrinos

Motivated by the recently shown connection between self-attention and (kernel) principal component analysis (PCA), we revisit the fundamentals of PCA. Using the difference-of-convex (DC) framework, we present several novel formulations and provide new theoretical insights. In particular, we show the kernelizability and out-of-sample applicability for a PCA-l

William Kuszmaul, Jingxun Liang, Renfei Zhou

Dynamic filters are data structures supporting approximate membership queries to a dynamic set $S$ of $n$ keys, allowing a small false-positive error rate $\varepsilon$, under insertions and deletions to the set $S$. Essentially all known constructions for dynamic filters use a technique known as fingerprinting. This technique, which was first introduced by

Giovanni Forni, Nelson Moll

We prove the existence of solutions of the cohomological equation for the geodesic flow on the unit tangent bundle of a compact flat surface with finitely many cone points. We also prove the ergodicity of the holonomy foliation for surfaces with non-rational holonomy, and the cohomology-free property of the horizontal foliated Laplacian under a simultaneous

Dharmik Patel, Antonio Rafael Vazquez Pantoja, Jiuzhou Lei, Kiju Lee

Fruit harvesting remains predominantly a labor-intensive process, motivating the development of research for robotic grippers. Conventional rigid or vacuum-driven grippers require complex mechanical design or high energy consumption. Current enveloping-based fruit harvesting grippers lack adaptability to fruits of different sizes. This paper introduces a dra

Nicola Bariletto, Stephen G. Walker

The consistency of posterior distributions in density estimation is at the core of Bayesian statistical theory. Classical work established sufficient conditions, typically combining KL support with complexity bounds on sieves of high prior mass, to guarantee consistency with respect to the Hellinger distance. Yet no systematic theory explains a widely held b

Mohamed Warda, Ronojoy Adhikari

The design and control of soft robots operating in fluid environments requires a careful understanding of the interplay between large elastic body deformations and hydrodynamic forces. Here we show that this interplay leads to novel elastohydrodynamic instabilities in a clamped soft robotic arm driven terminally by a constant pressure in a viscous fluid. We

Marco Pignatari, Roberto Gallino, Rene Reifarth

A clear definition of the contribution from the slow neutron-capture process (s process) to the solar abundances between Fe and the Sr-Zr region is a crucial challenge for nuclear astrophysics. Robust s-process predictions are necessary to disentangle the contribution from other stellar processes producing elements in the same mass region. Nuclear uncertaint

Xiangbo Gao, Tzu-Hsiang Lin, Ruojing Song, Yuheng Wu

Collaborative driving systems leverage vehicle-to-everything (V2X) communication across multiple agents to enhance driving safety and efficiency. Traditional V2X systems take raw sensor data, neural features, or perception results as communication media, which face persistent challenges, including high bandwidth demands, semantic loss, and interoperability i

Sudarshan Babu, Phillip Lo, Xiao Zhang, Aadi Srivastava

We introduce HyperDiffusionFields (HyDiF), a framework that models 3D molecular conformers as continuous fields rather than discrete atomic coordinates or graphs. At the core of our approach is the Molecular Directional Field (MDF), a vector field that maps any point in space to the direction of the nearest atom of a particular type. We represent MDFs using

Yonghao Zhuang, Junda Chen, Bo Pang, Yi Gu

We present core attention disaggregation (CAD), a technique that improves long-context large language model training by decoupling the core attention computation, softmax(QK^T)V, from the rest of the model and executing it on a separate pool of devices. In existing systems, core attention is colocated with other layers; at long context lengths, its quadratic