April 2026 arXiv papers — page 126

Francesco Andrea Causio, Vittorio De Vita, Olivia Riccomi, Michele Ferramola

While Large Language Models (LLMs) have demonstrated high proficiency on English-centric medical examinations, their performance often declines when faced with non-English languages and multimodal diagnostic tasks. This study protocol describes the development of EuropeMedQA, the first comprehensive, multilingual, and multimodal medical examination dataset s

Austin Talbot, Alex V. Kotlar, Yue Ke

Targeted amplicon panels are widely used in oncology diagnostics, but providing per-gene performance guarantees for copy number variant (CNV) detection remains challenging due to amplification artifacts, process-mismatch heterogeneity, and limited validation sample sizes. While Bayesian CNV callers naturally quantify per-sample uncertainty, translating this

Haoyu Hu, Yuelin Shao, Lorenzo Crippa, Dumitru Călugăru

Twisted bilayer graphene near integer fillings hosts correlated single-particle excitations whose dispersion and linewidth are increasingly accessible experimentally. We study these excitations using the topological heavy-fermion model, which captures both strong correlations and band topology of twisted bilayer graphene. In the decoupled limit, where both t

Arjun Bagchi, Saikat Mondal

We investigate fermions on Carrollian manifolds. We complement previous intrinsic analysis by deriving Carrollian fermion actions from a relativistic Dirac theory via a systematic expansion in the speed of light ($c$). We then study relativistic fermions in light-cone coordinates and their connection to Carrollian fermions in one lower dimension. This follow

Zhoutao Lei, Jihao Ma, Yun Chen, Tingting Wang

Exploiting enhanced sensitivity of a system in the vicinity of a phase transition boundary, critical quantum metrology to date still suffers from gap-closure related bottleneck effects, namely, critical slowing down of the sensing dynamics and a drastic shrinking of the parameter sensing window. To alleviate the said bottleneck inherent to any homogeneous la

Jun Zen, Xanda C. Kolesnikow, Campbell K. McLauchlan, Georgia M. Nixon

The ongoing development of hardware that is capable of reliably executing general quantum algorithms requires quantum error-correcting codes that are both practical for realisation and rapidly reduce logical error rates as they are scaled up. Here we introduce the dynamic compass code, a code that can be implemented with a modest footprint on the heavy-hex l

Anthony J. Brady, Yu-Xin Wang, Luis Pedro García-Pintos, Alexey V. Gorshkov

Measuring stochastic signals ("noise metrology") constitutes a central task in quantum sensing and the characterization of open quantum systems. Here we establish ultimate precision bounds for multiparameter estimation of stochastic signals encoded through Markovian Lindblad dynamics, allowing for arbitrary quantum control and noiseless ancillae. Although Ma

Yue Yu, Andrey V. Chubukov

In many strongly correlated electron systems, non-Fermi liquid behavior and unconventional superconductivity can be viewed as emerging from an effective 4-fermion interaction with a singular frequency dependence. A pairing instability in such a system is qualitatively different from that in a Fermi liquid and generally gives rise to multiple pairing states w

Andrey Chubukov, Ilya Esterlis, Artem Abanov, Nikolay Prokof'ev

The Migdal-Eliashberg theory (MET) describes electrons interacting with phonons in the adiabatic limit when the phonon Debye frequency is much smaller than the Fermi energy. A conventional belief is that MET holds even at strong coupling, when electron self-energy is large, and breaks down only near the point where the dressed phonon spectrum softens to near

Nikolay Prokof'ev, Ilya Esterlis, Artem Abanov, Andrey Chubukov

It is widely believed that in an adiabatic limit a Fermi liquid state of an electron-phonon system described by Migdal-Eliashberg theory remains stable before a dressed phonon softens. Using Holstein model as a prototypical example and variational/analytic considerations we demonstrate that in a wide range of fillings both in 3D and 2D, a polaronic/bi-polaro

H. R. Russell, P. E. J. Nulsen, A. C. Fabian, B. R. McNamara

We present a 100 ks XRISM Resolve observation of the Ophiuchus cluster that measures turbulence and bulk motion in the wake of the largest radio bubble on the sky. We detect a significant velocity shift of $-80\pm20$ km/s from the cluster centre to the bubble's wake and a clear increase in velocity dispersion from $135\pm10$ km/s to $210\pm20$ km/s. The meas

Dominik Nemeth, Ahsan Nazir, Robert-Jan Slager, Alessandro Principi

Boundary time crystals (BTCs) are prominent examples of continuous time crystals in collective spin systems governed by Lindbladian evolution. To date, their analysis has mostly relied on semiclassical and numerical approaches. Here, we develop a fully quantum-compatible framework to classify collective spin dynamics and show that BTC behavior emerges from t

Vasco Gennari, Tom Bertheas, Nicola Tamanini

Gravitational waves provide a powerful probe of both the astrophysical processes driving black hole mergers and the dynamics of the Universe, but these measurements rely on accurately inferring the unknown underlying population. We perform an agnostic reconstruction of the primary mass distribution using B-splines, characterising the emergence of structure w

Zhou-Quan Wan, Huan Jiang, Xuan Zou, Shiwei Zhang

Unlike conventional charge-2e superconductors, a charge-4e superconductor exhibits long-range coherence of electron quartets rather than Cooper pairs. Clear zero-temperature realizations of charge-4e superconductivity remain rare. Here, we investigate the zero-temperature phase diagram of the attractive SU(4) Hubbard model with numerically exact, large-scale

Guillermo Valverde, Igor García-Olaizola, Giannicola Scarpa, Alejandro Pozas-Kerstjens

Tensor networks were developed in the context of many-body physics as compressed representations of multiparticle quantum states. These representations mitigate the exponential complexity of many-body systems by capturing only the most relevant dependencies. Due to the formal similarity between quantum entanglement and statistical correlations, tensor networ

Sondre Duna Lundemo, Asle Sudbø

We show that the spin conductivity of a metal approaching a ferromagnetic quantum critical point exhibits divergent fluctuation corrections. This effect arises from critical spin fluctuations and constitutes a spin analog of the Aslamazov-Larkin theory of paraconductivity in superconductors. The spin current is derived in linear response within a Gaussian-le

Sarah Gaiser, Alessandro Russo, Philip Schuster

Axion-like particles (ALPs), the QCD axion, and dark photons in the MeV-GeV mass range are motivated by various dark matter models and the strong CP problem, and are ubiquitous in extensions of the Standard Model. A long-standing blind spot for experimental searches is the sub-100 MeV mass range, where the particle lifetime is too long to be constrained by p

Peiran Li, Zhen Liu, Lian-Tao Wang

Higgs final states are prime targets in the search for physics beyond the Standard Model. In the conventional picture, $SU(2)$ symmetry together with the Goldstone Equivalence Theorem correlates Higgs and gauge-boson final states, implying comparable sensitivity in channels such as $hh$, $ZZ$, and $WW$ in searches for heavy resonances. In this work, we ident

Martin Beneke, Patrick Hager, Andrea F. Sanfilippo

In the framework of Soft de Sitter Effective Theory (SdSET), the Fokker-Planck equation for the late-time dynamics of the massless minimally coupled scalar field and its extension to the Kramers-Moyal equation are obtained from operator mixing of composite operators of the effective superhorizon field. We construct the formalism for composite-operator renorm

Matteo Borrello, Marco Costa, Diego Redigolo, Michele Tammaro

We explore the possibility that neutrinos couple to an interacting sterile sector, providing a novel portal that generalizes the heavy neutral lepton portal to a composite setting. For a low confinement scale, high-energy neutrino beams can disintegrate into collimated sprays of hidden states, referred to as dark jets. This dynamics gives rise to two charact

Dario Buttazzo, Gabriele Levati, Yang Ma, Fabio Maltoni

The electric and magnetic dipole moments of the electron and of the muon provide stringent tests of the Standard Model and sensitive probes of new physics. By contrast, the corresponding dipole moments of the $\tau$ lepton remain weakly constrained. This study explores the potential of future lepton colliders, focusing on the $e^+e^-$ Future Circular Collide

Johannes Stoiber, Lucas M. Valenzuela, Rhea-Silvia Remus, Klaus Dolag

Tidal stellar shells and streams are two of the most intriguing low-surface-brightness features within galaxies, consisting of stars accreted from satellite galaxies. A crucial ingredient in determining which type of feature will be formed is the orbit of the satellite galaxy. Additionally, the distribution of stars from these satellite galaxies within the m

Aurélien Dersy, Matthew D. Schwartz

The path-integral approach to the double well has long been limited by the dilute instanton gas approximation. We show that if the finite Euclidean-time structure is taken seriously by using exact saddles, the dilute gas can be sidestepped, allowing the partition function and energy levels to be computed systematically. At each instanton order, the full resu

Yaar Vituri, Erez Berg

We develop a controlled theoretical framework for the topological heavy fermion model relevant to magic-angle twisted bilayer graphene, where low density conduction electrons hybridize with a lattice of strongly interacting f-sites. By tracing out the localized electrons, we derive an effective action for the conduction electrons with long-range in time effe

Laura Shou, Joseph T. Iosue, Yu-Xin Wang, Victor Galitski

We study the growth of entanglement and circuit complexity in random passive linear optical networks as a function of the circuit depth. For entanglement dynamics, we start with an initial Gaussian state with all $n$ modes squeezed. For random brickwall circuits, we show that entanglement, as measured by the R\'enyi-2 entropy, grows at most diffusively as a

Mona El Morsy, Olivier Guyon, Barnaby Norris, Sergio Leon-Saval

HWO aims to directly image objects orbiting Sun-like stars, using a 6-m telescope capable of high-contrast imaging ($10^{-10}$) and spectroscopy to search for biosignatures in planets located in the habitable zone. Recent laboratory demonstrations and ground-based telescope projects have shown the effectiveness of SMFs in spectroscopy, paving the way for SMF

Jonathan J. Heckman, Rebecca J. Hicks, Chitraang Murdia

Non-invertible symmetries of a quantum field theory (QFT) are a natural generalization of unitary symmetries, but in which the product of operators does not satisfy a group multiplication law. We show that such symmetry operations on states define a collection of quantum gates for a parallel quantum computation scheme that includes post-selection / projectio

Julian L. Northey, Yael Shadmi, Yotam Soreq, Daiki Ueda

We use the amplitude formulation of the SMEFT to introduce a spurion analysis of the SMEFT low-energy amplitudes in terms of the Higgs VEV. Each SMEFT contact-term is given as a sum of a few spurion structures, whose number depends on the electroweak charges of the external legs. The coefficients of these structures involve singlet combinations of Higgses fr

Lucas M. Valenzuela, Johannes Stoiber, Rhea-Silvia Remus

The combination of deep imaging data and kinematic measurements in galaxy outskirts promises to reveal extensive insights into the structure and history of individual galaxies. From a census of tidal features around galaxies from the Magneticum simulation, we disentangle the dynamics for a selected stellar stream from the underlying halo by identifying the s

William Munizzi, Howard J. Schnitzer

Non-Clifford gates, used to generate quantum magic, are essential for universal quantum computation. We show that non-Clifford gates arise naturally from path integrals in topological quantum field theories, where their magic-generating properties are determined by the algebraic data of the theory. In Chern-Simons theory, we construct the Ising interaction g

Christopher Whittall, Geraint Pratten

Gravitational waveform templates are a key ingredient for the detection and characterization of gravitational waves emitted by compact binary mergers in the universe. These templates must be physically accurate and extensive, but also highly computationally efficient, two requirements that are often in tension. One solution to this problem is the development

Emanuel Gafton, Illa R. Losada

We present Visplot, a free and open-source tool designed for interactive construction and refinement of nightly telescope observing schedules through a graphical user interface. The software performs hardware-aware observability analysis by explicitly incorporating telescope-specific limits alongside astronomical constraints, in order to determine feasible o

Juan Pablo Esparza, Francisco J. Peña, Patricio Vargas, Vladimir Juričić

Non-Hermitian band descriptions capture how loss, gain, and environmental coupling reshape quantum matter, yet most experimental tests rely on wave-based or dynamical probes. Here we establish a new equilibrium route to exceptional physics in Dirac materials: in the weakly non-Hermitian regime, the thermodynamic density of states and the quantum capacitance

Zheyu Zhang, Ziqi Pang, Shixing Chen, Xiang Hao

Long video understanding is inherently challenging for vision-language models (VLMs) because of the extensive number of frames. With each video frame typically expanding into tens or hundreds of tokens, the limited context length of large language models (LLMs) forces the VLMs to perceive the frames sparsely and lose temporal information. To address this, we

Team Seedance, De Chen, Liyang Chen, Xin Chen

Seedance 2.0 is a new native multi-modal audio-video generation model, officially released in China in early February 2026. Compared with its predecessors, Seedance 1.0 and 1.5 Pro, Seedance 2.0 adopts a unified, highly efficient, and large-scale architecture for multi-modal audio-video joint generation. This allows it to support four input modalities: text,

Song Tang, Guangquan Jie, Henghui Ding, Yu-Gang Jiang

Existing segmentation models based on multimodal large language models (MLLMs), such as LISA, often struggle with novel or emerging entities due to their inability to incorporate up-to-date knowledge. To address this challenge, we introduce the Novel Emerging Segmentation Task (NEST), which focuses on segmenting (i) novel entities that MLLMs fail to recogniz

Konstantinos Ladovrechis

The interplay among topology and interactions has been a promising path towards identifying novel phases of condensed matter beyond these predicted by the established classification paradigms. In the present work, we propose such a novel phase of matter by studying the fate of a three-dimensional Berry-dipole semimetal, lying at the topological quantum criti

Team HY-World, Chenjie Cao, Xuhui Zuo, Zhenwei Wang

We introduce HY-World 2.0, a multi-modal world model framework that advances our prior project HY-World 1.0. HY-World 2.0 accommodates diverse input modalities, including text prompts, single-view images, multi-view images, and videos, and produces 3D world representations. With text or single-view image inputs, the model performs world generation, synthesiz

Dinging Li, Yingxiu Zhao, Xinrui Cheng, Kangheng Lin

Spatial reasoning over three-dimensional scenes is a core capability for embodied intelligence, yet continuous model improvement remains bottlenecked by the cost of geometric annotation. The self-evolving paradigm offers a promising path, but its reliance on model consensus to construct pseudo-labels causes training to reinforce rather than correct the model

Samuel J. Garratt, Dmitry A. Abanin

We construct upper bounds on entanglement entropies of many-body quantum states that have fixed energy expectation values with respect to geometrically local Hamiltonians. Our focus is on entanglement entropies of subsystems that make up approximately half of the full system. The upper bound on the von Neumann entanglement entropy is half the sum of the ther

Yuqiao Tan, Minzheng Wang, Bo Liu, Zichen Liu

While reinforcement learning with verifiable rewards (RLVR) significantly enhances LLM reasoning by optimizing the conditional distribution P(y|x), its potential is fundamentally bounded by the base model's existing output distribution. Optimizing the marginal distribution P(y) in the Pre-train Space addresses this bottleneck by encoding reasoning ability an

Lin-Zhuo Chen, Jian Gao, Yihang Chen, Ka Leong Cheng

Streaming 3D reconstruction aims to recover 3D information, such as camera poses and point clouds, from a video stream, which necessitates geometric accuracy, temporal consistency, and computational efficiency. Motivated by the principles of Simultaneous Localization and Mapping (SLAM), we introduce LingBot-Map, a feed-forward 3D foundation model for reconst

Sumeet Ramesh Motwani, Daniel Nichols, Charles London, Peggy Li

As language models are increasingly deployed for complex autonomous tasks, their ability to reason accurately over longer horizons becomes critical. An essential component of this ability is planning and managing a long, complex chain-of-thought (CoT). We introduce LongCoT, a scalable benchmark of 2,500 expert-designed problems spanning chemistry, mathematic

Raphael Tsiamis

For every closed set $K \subset \mathbb{R}^n$ and every $m \geq 2$, we construct a mean-convex ancient solution to mean curvature flow of hypersurfaces in $\mathbb{R}^{m+n}$, with respect to a smooth Riemannian metric arbitrarily $C^\infty$-close to the Euclidean metric, whose first-time singular set is exactly $K \times \{0\}$.

Alessandra Caraceni, Nicolas Curien, William Fleurat, Adrianus Twigt

We present a simple trimming algorithm that generates nested uniform binary plane trees by removing leaves one-by-one using a best-of-three-match procedure. While its one-step transition specializes to the Luczak-Winkler & Caraceni-Stauffer coupling, its scaling limit provides a suprising c\`adl\`ag erasure of Brownian trees, reminiscent of SLE theory.

Grace Krahm, Adam K. Leroy, Jiayi Sun, Kijeong Yim

We present high-resolution (0.94" $\approx$ 55 pc) ALMA CO(2-1) and 13CO(2-1) observations of the highly inclined (i~87.5 deg) galaxy NGC 4565 covering out to galactocentric radius Rgal > $\pm$ 17 kpc. The combination of sensitivity and resolution enables the detection of CO emission well into the HI-dominated outer disk while isolating individual molecular

Giacomo Brunello, Manoj K. Mandal, Pierpaolo Mastrolia, Raj Patil

We present a closed formula for the computation of static post-Newtonian corrections to the two-body gravitational dynamics at any odd order, assuming the lower-order results are known. The formula is derived within a correlation function framework and exploits the $\mathbb{Z}_2$ symmetry of the static sector, leading to a novel theoretical interpretation of

Zhong-Bo Kang, Luke Sellers, Congyue Zhang, Curtis Zhou

We present an extraction of unpolarized quark transverse-momentum-dependent parton distribution functions (TMD PDFs) from Drell-Yan data within a Bayesian inference framework, incorporating artificial intelligence at multiple stages of the analysis. Our analysis is performed at ${\rm N^3LO}$ in perturbative QCD combined with ${\rm N^4LL}$ resummation accurac

Kajol Paithankar, Sanved Kolekar

Black hole shadow images are primarily determined by the properties of photon spheres and can exhibit degeneracies across different spherically symmetric spacetime geometries. We show that time delay observables associated with higher-order images of transient sources provide a robust probe to break such degeneracies in spacetimes admitting multiple photon s

Yu-Qi Lei, Xian-Hui Ge, Yu Tian, Shao-Feng Wu

Understanding how temporal order degrades in quantum systems remains a central issue in nonequilibrium physics. Here we study the melting of discrete time crystals in a periodically driven holographic system, where a distinct (discrete) time semicrystal phase emerges with persistent temporal order in disorder, bridging discrete time crystals and fully disord

Yang Hu, Na Li

In this paper, we propose a novel framework for the joint identification of system dynamics and noise covariance in linear systems, under general noise distributions beyond Gaussian. Specifically, we would like to simultaneously estimate the dynamical matrix $A$ and the noise covariance matrix $\varSigma$ using state transition data. The formulation builds u

Ami Baid, Zihui Xue, Kristen Grauman

While Audio-Visual Language Models (AVLMs) have achieved remarkable progress over recent years, their reliability is bottlenecked by cross-modal hallucination. A particularly pervasive manifestation is video-driven audio hallucination: models routinely exploit visual shortcuts to hallucinate expected sounds, discarding true auditory evidence. To counteract t

Louie Hong Yao, Vishesh Anand, Yuan Zhuang, Tianyu Jiang

Rhetorical questions are asked not to seek information but to persuade or signal stance. How large language models internally represent them remains unclear. We analyze rhetorical questions in LLM representations using linear probes on two social-media datasets with different discourse contexts, and find that rhetorical signals emerge early and are most stab

Panagiotis Dimakis, Duong Dinh, Shengjing Xu

On a compact connected Riemann surface $C$ of genus at least $2$, we construct Lagrangian correspondences between moduli spaces of rank-$n$ Higgs bundles (respectively, holomorphic connections) and the Hilbert schemes of points on $T^\ast C$ (respectively, the twisted cotangent bundles of $C$). Central to these constructions are Higgs bundles (respectively,

Erjia Yan, Chaoqun Ni

Generative AI systems such as ChatGPT are increasingly used in scientific writing, yet their broader implications for the organization of scientific knowledge remain unclear. We examine whether AI-assisted writing intensity, measured as the share of text in a paper that is predicted to exhibit features consistent with LLM-generated text, is associated with s

Tianshuo Yang, Guanyu Chen, Yutian Chen, Zhixuan Liang

While end-to-end Vision-Language-Action (VLA) models offer a promising paradigm for robotic manipulation, fine-tuning them on narrow control data often compromises the profound reasoning capabilities inherited from their base Vision-Language Models (VLMs). To resolve this fundamental trade-off, we propose HiVLA, a visual-grounded-centric hierarchical framewo

Nicolas De Ro, Jonathan Demaeyer, Stéphane Vannitsem

Baroclinic instability is a fundamental mechanism driving atmospheric dynamics. In this work, we revisit Pedlosky's two-layer model for finite amplitude baroclinic waves - a seminal framework for studying the unstable growth of finite perturbations - leveraging modern nonlinear techniques and computational resources. We show that the geophysical state of the

Irina Đanković, Maarten Markering, Jason Miller, Yizheng Yuan

We consider critical site percolation ($p=p_c=1/2$) on the triangular lattice $\mathbf{T}$ in two dimensions. We show that the simple random walk on the clusters of open vertices converges in the scaling limit to a continuous diffusion which lives in the gasket of a conformal loop ensemble with parameter $\kappa = 6$ $\big(\mathrm{CLE}_6\big)$, the so-called

Zipeng Ling, Shuliang Liu, Seonil Son, Shenghong Fu

Large language models (LLMs) have become increasingly used for various tasks, often coupled with Chain-of-Thought (CoT) prompting to boost accuracy. Recent work has shown that high label-prediction accuracy does not guarantee correct intermediate reasoning, and the causes of *reasoning flaws* vary from sample to sample, yet existing remedies either focus on

Nathaniel Hai, Chinmayee Vallabh Prabhu Dessai, Dingcheng Sun, Jianpeng Ao

Intracellular chromophores (e.g., NADH and FAD) play a central role in regulation of cellular metabolism. Though autofluorescence has been extensively used for label-free mapping of chromophores inside a cell, its sensitivity and molecular specificity are constrained by the low quantum yield and the fluorescence spectral overlap. Here, we address these chall

Khen Cohen, Yoav Yosif-Or, Yaron Oz, Ady Arie

A mode sorter separates a set of M orthogonal spatial modes in a shared input channel into M different output channels. Here we present an analytic derivation and experimental validation of a single plane device for sorting spatial modes from a diverse variety of mode families, including Hermite-Gaussian (HG), Laguerre-Gaussian (LG), Bessel-Gaussian (BG), wi

Adi Arbel, Stefan Steinerberger, Ronen Talmon

Given two symmetric positive-definite matrices $A, B \in \mathbb{R}^{n \times n}$, we study the spectral properties of the interpolation $A^{1-x} B^x$ for $0 \leq x \leq 1$. The presence of `common structures' in $A$ and $B$, eigenvectors pointing in a similar direction, can be investigated using this interpolation perspective. Generically, exact log-lineari

Alexander Adam, Szabolcs Borsányi, Zoltán Fodor, Jana N. Guenther

Quantum Chromodynamics (QCD) at finite density is most often formulated on the lattice as a grand canonical ensemble. Since lattice QCD has a complex action problem at finite baryo-chemical potential ($\mu_B$), its results at finite density are indirect: e.g. in the form of a set of expansion coefficients. In contrast, the canonical formulation offers direct

Zerun Ma, Guoqiang Wang, Xinchen Xie, Yicheng Chen

While Large Language Models (LLMs) have empowered AI research agents to perform isolated scientific tasks, automating complex, real-world workflows, such as LLM training, remains a significant challenge. In this paper, we introduce TREX, a multi-agent system that automates the entire LLM training life-cycle. By orchestrating collaboration between two core mo

Junzhe Wang, Zhiheng Xi, Yajie Yang, Hao Luo

Search agents extend Large Language Models (LLMs) beyond static parametric knowledge by enabling access to up-to-date and long-tail information unavailable during pretraining. While reinforcement learning has been widely adopted for training such agents, existing approaches face key limitations: process supervision often suffers from unstable value estimatio

Airi Kawasaki, Fei Gao, Gustavo E. Scuseria

In this paper, we develop a class of antisymmetrized geminal power configuration interaction (AGP-CI) wave functions that extend the AGP framework by incorporating inter-geminal correlations through a CI expansion. To make these wavefunctions computationally tractable, we evaluate them by rewriting the AGP-CI ansatz as a linear combination of AGPs (LC-AGP),

Xiaofan Zhou, Kyumin Lee

Sequential recommendation has become increasingly prominent in both academia and industry, particularly in e-commerce. The primary goal is to extract user preferences from historical interaction sequences and predict items a user is likely to engage with next. Recent advances have leveraged contrastive learning and graph neural networks to learn more express

Tadhg Looram, Lucas Nuzzi, Kyle Waters, Steven Dillmann

Much of the focus in RL today is on evaluation design: building meaningful evals that serve simultaneously as benchmarks and as well-defined reward signals for post-training. Yet, many real-world tasks are governed by subjective, procedural, and domain-specific requirements that are difficult to encode as exact-match targets or open-ended preference judgment

Fei Tang, Bofan Chen, Zhengxi Lu, Tongbo Chen

GUI grounding, which localizes interface elements from screenshots given natural language queries, remains challenging for small icons and dense layouts. Test-time zoom-in methods improve localization by cropping and re-running inference at higher resolution, but apply cropping uniformly across all instances with fixed crop sizes, ignoring whether the model

Chih-Hao Chang, Chia-Hsien Shen, Zihan Zhou

In the effective field theory (EFT) description of binary inspirals, the radiated gravitational waveform receives universal corrections from the curved background, the so-called ``tail effects'', that resum into the so-called ``Sommerfeld factor''. We develop a systematic framework for computing this gravitational Sommerfeld factor for scalar perturbations w

Swati Rallapalli, Shannon Gallagher, Ronald Yurko, Tyler Brooks

Large Language Models (LLMs) are now capable of generating highly fluent, human-like text. They enable many applications, but also raise concerns such as large scale spam, phishing, or academic misuse. While much work has focused on detecting LLM-generated text, only limited work has gone into understanding the stylistic differences between human-written and

Xin An, Gokce Basar, Mikhail Stephanov

We derive deterministic equations for the evolution of non-Gaussian fluctuations in relativistic stochastic hydrodynamics. This is achieved by defining the average local Landau frame and corresponding fluctuating hydrodynamic variables. Fully nonlinear stochastic hydrodynamics is expressed in a unified multi-component matrix form. A novel relativistic formal

Arseniy Andreyev, Advikar Ananthkumar, Marc Walden, Tomaso Poggio

Recent work suggests that (stochastic) gradient descent self-organizes near an instability boundary, shaping both optimization and the solutions found. Momentum and mini-batch gradients are widely used in practical deep learning optimization, but it remains unclear whether they operate in a comparable regime of instability. We demonstrate that SGD with momen

Arnob Barua, Christopher E. Kees, Dmitri Kuzmin

Simulating infiltration in porous media using Richards' equation remains computationally challenging due to its parabolic structure and nonlinear coefficients. While a wide range of numerical methods for differential equations have been applied over the past several decades, basic higher-order numerical methods often fail to preserve physical bounds on water

Jordan Taylor, Joel Mire, Alicia DeVrio, Maarten Sap

Art-making is a collective social activity through which queer people engage in political resistance, develop identities, archive queer memory, and form community. However, in recent years, generative AI has disrupted queer artistic communities. Through 15 semi-structured interviews, we examine how queer artists are making sense of the encroachment of GenAI

David Gao, Srivatsav Kunnawalkam Elayavalli

We prove a new "Toeplitz exactness" theorem for strong convergence. This is a machine to upgrade strong convergence in the general setting of $C^\ast$-correspondences, and has several applications.

Aubril Ony

We give explicit axioms for the algebraic theory of the quasivarieties of right-preordered groups and preordered groups. We then look at lattices of effective equivalence relations, which turn out to be similar to the lattices of equivalence relations in the category of groups. Once this is established, we study internal structures in the category of right-p

H. P. Pinheiro, C. A. S. Almeida

The localization of chiral fermionic zero modes in scalar field backgrounds with domain wall structure is a central mechanism in brane-world scenarios. We investigate this mechanism in a system that provides an effective realization of the $(1+1)$-dimensional Jackiw--Rebbi model, using a two-kink scalar background generated by the deformation method applied

Roni Ben-Maimon, Ofer Firstenberg, Nir Davidson, Ephraim Shahmoon

We study quantum light-matter interfaces based on bilayer atomic arrays in free space, considering interlayer spacings $a_z$ that may deviate from the Bragg-symmetric condition, $a_z\in \mathrm{integer}\times \lambda/2$ with $\lambda$ the light wavelength. Mapping the problem to a one-dimensional model, we show that the interface efficiency is fully determin

Lucio Galeati

In this note we show the existence of a residual set (in the sense of Baire) of divergence free initial data $u_0\in L^2(D)$, $D=\mathbb{R}^2$ or $\mathbb{T}^2$, for which global existence and uniqueness of weak solutions to the incompressible 2D Euler equations holds. The associated solutions $u$ satisfy the energy balance and are recovered in the vanishing

Angela Conaci, Stefania De Curtis, Luigi Delle Rose, Atri Dey

We show that the Aligned 2-Higgs Doublet Model (A2HDM) is a framework able to simultaneously accommodate strong first order electro-weak phase transitions, in turn generating detectable gravitational waves as well as a variety of Higgs boson signals (involving both the Standard Model state and its companions, both neutral and charged) accessible at the Large

Wojciech Gorecki, Christiane P. Koch

Quantum metrology promises precision beyond classical limits but environmental noise, unless properly controlled, reduces the quantum advantage to at most a constant improvement. A key challenge is therefore to design quantum control strategies that suppress noise while preserving sensitivity to the targeted signal. Here, we suggest to use dressed states gen

Radwa Sultan

Designing robust architectures that can mitigate sophisticated attacks is now a key priority for modern wireless systems. This paper investigates a single-cell bistatic integrated sensing and communication (ISAC) network facing simultaneous coordinated active jamming and malicious detection. These threats aim to disrupt the downlink communication and detect

Haoran Xu, Kaiwen Hu, Somayeh Sojoudi, Amy Zhang

We study behavior-regularized reinforcement learning (RL), where regularization toward a reference distribution (the dataset in offline RL or the base model in LLM RL finetuning) is essential to prevent value over-optimization caused by erroneous out-of-distribution extrapolation. Existing methods either rely on reparameterized policy gradient, which are dif

Ines Bentara

Star evolution models predict the lightest compact objects in the universe to have masses greater than that of the Sun. Nonetheless, alternative scenarios could lead to the formation of sub-solar mass (SSM) compact objects, such as primordial black holes (PBHs). The LIGO-Virgo-KAGRA Collaboration (LVK) has performed a search for gravitational-wave (GW) signa

Min Huang, Yandong Liu, Hao Zhang

In this work, we investigate three typical new physics resonances which couple to the standard model (SM) quarks via direct top-quark flavor-changing interactions. We identify the possible SMEFT operators electroweak scale and analyze their phenomenology.

Simon Ostermann, Daniil Gurgurov, Tanja Baeumel, Michael A. Hedderich

Post-training adaptation of language models is commonly achieved through parameter updates or input-based methods such as fine-tuning, parameter-efficient adaptation, and prompting. In parallel, a growing body of work modifies internal activations at inference time to influence model behavior, an approach known as steering. Despite increasing use, steering i

Ziming Wang

We present UMI-3D, a multimodal extension of the Universal Manipulation Interface (UMI) for robust and scalable data collection in embodied manipulation. While UMI enables portable, wrist-mounted data acquisition, its reliance on monocular visual SLAM makes it vulnerable to occlusions, dynamic scenes, and tracking failures, limiting its applicability in real

Alessandro Carones, Sijil Jose, Aliza Mustafa, Nicoletta Krachmalnicoff

We present BROOM, a new python package for the application of blind, minimum-variance component-separation techniques to microwave observations. The package enables the reconstruction of signals with known spectral energy distributions, such as the Cosmic Microwave Background (CMB), Sunyaev--Zeldovich distortions, or foreground moments, in both temperature a

Liam Mazurowski, Xuan Yao

We address Gromov's Quantification of $C^0$ Convergence Conjecture in dimension three. Let $B$ be the unit ball in $\mathbb R^3$. Let $g$ and $g_0$ be smooth metrics on $B$. We prove there are constants $C$ and $\epsilon_0$ depending only on $g_0$ so that \[ \inf_{x\in B} R_g(x) \leq R_{g_0}(0) + C \|g-g_0\|_{C^0}^{1/2} \] provided $\|g-g_0\|_{C^0}\leq \epsi

Qi Zhang

In the field of online reinforcement learning (RL), traditional Gaussian policies and flow-based methods are often constrained by their unimodal expressiveness, complex gradient clipping, or stringent trust-region requirements. Moreover, they all rely on post-hoc penalization of negative samples to correct erroneous actions. This paper introduces Positive-On

Harsh Parikh, Tyler McCormick, Emily Johnson, Leo Hickey

Artificial intelligence (AI) systems increasingly shape how people access health information, make medical decisions, and receive care -- yet epidemiology lacks frameworks for measuring AI exposure or studying its health effects at the population level. Here we argue that AI now functions as a determinant of health and propose a conceptual framework, borrowe

Jose Recatala-Gomez, Haiwen Dai, Zhu Ruiming, Nikita Kazeev

Materials discovery is fundamental to advance next-generation technologies as well as for sustainable and circular economy. Beyond computational screening, generative models are efficient at finding materials with desired properties, via multi-modal learning using multiscale data. This perspective examines the landscape of generative design for inorganic mat

Huaijing Huang, Daowen Qiu, Ximing Hua, Xinyu Chen

Grover's algorithm accelerates unstructured database search quadratically compared to classical algorithms. In the NISQ era, distributed quantum computing can decrease circuit depth and reduce noise. In this paper, an algorithm for constructing query operators for subfunctions is proposed. By dividing the target string of the search problem into several subs

Yong Zhang, Nicolas Labrosse, Sargam M. Mulay

An observation of a large prominence on the solar limb took place on February 15, 2022, by the Extreme Ultraviolet Imager (EUI) on board Solar Orbiter. We aim to determine the range of physical parameters of this prominence, such as temperature, radial velocity, and altitude, and examine how these parameters affect the formation of the 304~\AA\ line of \ion{

Shi Jin, Nana Liu, Chuwen Ma

We propose quantum algorithms for complex-valued nonlinear partial differential equations in the strongly nonlinear regime, where the dynamics is governed by vortex cores, phase singularities, and nonlinear vortex interactions. Examples include the complex-valued nonlinear Schr\"odinger equation, as well as nonlinear heat and wave equations with Ginzburg--La

Sajid Mannan, K. Sidharth Nambudiripad, Indrajeet Mandal, Nitya Nand Gosvami

Long-term chemical durability of glass, crucial for immobilizing nuclear waste, is governed by glass properties such as composition, surface geometry, as well as external factors like thermodynamic conditions and surrounding medium. Despite decades of research, there are no models that account for these intrinsic and extrinsic factors to predict the dissolut

Behnaz Elhaminia, Catherine King, Jiaqi Lv, Lorraine Harper

Accurate detection and segmentation of glomeruli in kidney tissue are essential for diagnostic applications. Traditional deep learning methods primarily rely on semantic segmentation, which often fails to precisely delineate adjacent glomeruli. To address this challenge, we propose a novel glomerulus detection and segmentation model that emphasises boundary

Alessandro Proserpio, Ian A. B. Strachan

The idea of a $\bigvee$-system was introduced by Veselov in the study of rational solutions of the WDVV equations of associativity. These are algebraic/geometric conditions on the set of covectors that appear in rational solutions to the WDVV equations. Here, this idea is generalized to open WDVV equations, which are an additional set of PDEs originating fro

Joseph Klobusicky, Matthew Rakauskas

We present a model for sticky particles in which cluster sizes after a reaction have $\ell$ fewer total particles than the sum of their reactants. The finite particle system is modeled as a Markov process under a mean-field assumption for selecting reactants. The limiting kinetic equations form an infinite system of nonlinear differential equations similar t