March 2026 arXiv papers — page 129

Baleegh Abdo, William Shanks, Oblesh Jinka, J. R. Rozen

The ability to perform high-fidelity quantum nondemolition qubit readout is pivotal for the realization of large and powerful quantum computers. Such readout of superconducting qubits is generally enabled by amplifying the weak dispersive measurement signals using phase-preserving quantum-limited Josephson amplifiers with sufficient gain to dilute the contri

Linrui Ma, Yufei Cui, Kai Han, Yunhe Wang

Discrete diffusion models offer global context awareness and flexible parallel generation. However, uniform random noise schedulers in standard DLLM training overlook the highly non-uniform information density inherent in real-world sequences. This wastes optimization resources on low-density structural glues while leaving high-density logical pivot points s

Ce Zhang, Jinxi He, Junyi He, Katia Sycara

Multi-modal Large Language Models (MLLMs) have achieved remarkable performance across a wide range of visual reasoning tasks, yet their vulnerability to safety risks remains a pressing concern. While prior research primarily focuses on jailbreak defenses that detect and refuse explicitly unsafe inputs, such approaches often overlook contextual safety, which

Vatsal Gupta, Darshan Sreenivasamurthy

Prose2Policy (P2P) is a LLM-based practical tool that translates natural-language access control policies (NLACPs) into executable Rego code (the policy language of Open Policy Agent, OPA). It provides a modular, end-to-end pipeline that performs policy detection, component extraction, schema validation, linting, compilation, automatic test generation and ex

Alexandre Lacoste, Nicolas Gontier, Oleh Shliazhko, Aman Jaiswal

The proliferation of agent benchmarks has created critical fragmentation that threatens research productivity. Each new benchmark requires substantial custom integration, creating an "integration tax" that limits comprehensive evaluation. We propose CUBE (Common Unified Benchmark Environments), a universal protocol standard built on MCP and Gym that allows b

Hao Wu, Yongheng Zhang, Yuan Gao, Fan Xu

Large Language Models (LLMs) have demonstrated exceptional logical reasoning capabilities but frequently struggle with the continuous spatiotemporal dynamics governed by Partial Differential Equations (PDEs), often resulting in non-physical hallucinations. Existing approaches typically resort to costly, domain-specific fine-tuning, which severely limits cros

Eric Penner, Josephine D'Angelo, Clinton Smith, Nathan Matsuda

End-to-end (e2e) latency in head-mounted displays (HMD) is the time delay between a physical change in the world (e.g., a user's head movement) and the moment the display updates to reflect that change. Tracking, rendering, and other computation in real systems invariably introduce some amount of e2e latency to all HMDs. In modern devices this latency is usu

Tuoping Du, Zhifeng Peng

This paper establishes an arithmetic intersection formula for central L-derivatives in higher weights.We prove that for a general cusp form (extending the previous result for newforms), the derivative is represented by the global height pairing between higher Heegner cycles. This result provides a framework for the Gross-Zagier-Zhang formula and its generali

Francisco Albarrán-Arriagada, Juan Carlos Retamal

Adiabatic ground-state preparation is fundamentally limited by the spectral structure of the time-dependent Hamiltonian, particularly by gap reductions and degeneracies that induce nonadiabatic transitions. We examine this dependence in the anisotropic Heisenberg (XXZ) model on an eight-site ring by comparing three strategies: optimization of the initial Ham

H. Miao, G. Fabbris, J. Bouaziz, W. R. Meier

Chirality is a fundamental organizing principle of correlated and topological states. In quantum magnets, chirality arises from the geometric twisting of spins and serves as an emergent source of Berry curvature and quantum metrics. Although external fields can reversibly tune the spin chirality, understanding how spontaneous reversal occurs on macroscopic l

Giulia Mazzola, David Sutter, Renato Renner

Information-theoretic techniques are based on the assumption that resources are well characterized by independent and identically distributed (iid) states. This assumption cannot be justified operationally, since, for example, correlations between subsequent systems emitted by a source cannot be detected by any practical tomographic protocol. Operationally m

Sadaf Tabatabaee, Sarah S. Lam

Efficient classification of surgical procedures by urgency is paramount to optimize patient care and resource allocation within healthcare systems. This study introduces an unsupervised neural network approach to automatically categorize surgical transcriptions into three urgency levels: immediate, urgent, and elective. Leveraging BioClinicalBERT, a domain-s

Rachid Ouyed

We show that delayed (weeks-months) energy injection into expanding Type Ic supernova (SN) ejecta can reproduce the luminosity and spectral evolution of hydrogen-poor superluminous SNe (SLSNe-I). Late-time reheating sets the radiation temperature and density needed for the W-shaped OII absorption near peak, explaining its disappearance as the ejecta cools wi

Debanand Sa, Anirban Dutta

We have developed a semi-analytical framework formulated in the canonical fermion representation to investigate strongly correlated electron systems. We consider the U=$\infty$ Hubbard model and used the equation of motion method to calculate the fermion self-energy which has two parts: single and two-boson exchange processes. The emergent bosons here are se

Patrick Yin, Tyler Westenbroek, Zhengyu Zhang, Joshua Tran

Reinforcement learning in massively parallel physics simulations has driven major progress in sim-to-real robot learning. However, current approaches remain brittle and task-specific, relying on extensive per-task engineering to design rewards, curricula, and demonstrations. Even with this engineering, they often fail on long-horizon, contact-rich manipulati

Roan Talbut, Andrew McCormack, Anthea Monod

Fr\'echet means are a popular type of average for non-Euclidean datasets, defined as those points which minimise the average squared distance to a set of data points. We consider the behaviour of sample Fr\'echet means on normed spaces whose unit ball is a polytope; this setting is rarely covered by existing literature on Fr\'echet means, which focuses on sm

Yue Chen, Kyoung Ho Kim, Steven K. Morley, Jesse R. Woodroffe

This report presents a brief review of the 2024 Mother's Day superstorm and its impacts on the near-Earth space environment and the ground-level effects, with emphasis on the latter. Drawing upon observations from multiple sources. we qualitatively illustrate how intense space weather disturbances can generate strong geoelectric fields and drive pronounced g

Saba Asaad, Hina Tabassum, Ping Wang

This paper introduces a novel multi-objective integrated sensing and communications (ISAC) framework to enable collaborative wireless sensing in conjunction with over-the-air federated-edge learning (OTA-FEEL). The framework enables multi-task OTA aggregation to handle sensing and learning simultaneously, while benefiting from dual-purpose uplink signals for

Vladimir Kolmogorov, Jack Spalding-Jamieson

We present a family of fast pseudo-approximation algorithms for the minimum balanced vertex separator problem in a graph. Given a graph $G=(V,E)$ with $n$ vertices and $m$ edges, and a (constant) balance parameter $c\in(0,1/2)$, where $G$ has some (unknown) $c$-balanced vertex separator of size ${\rm OPT}_c$, we give a (Monte-Carlo randomized) algorithm runn

Nicolas A. Errandonea, Santiago Mazuelas, Jose A. Lozano, Sanjoy Dasgupta

Prior work on partial labels learning (PLL) has shown that learning is possible even when each instance is associated with a bag of labels, rather than a single accurate but costly label. However, the necessary conditions for learning with partial labels remain unclear, and existing PLL methods are effective only in specific scenarios. In this work, we mathe

Hippolyte Verninas, Caner Korkmaz, Stefanos Zafeiriou, Tolga Birdal

Machine learning has been progressively generalised to operate within non-Euclidean domains, but geometrically accurate methods for learning on surfaces are still falling behind. The lack of closed-form Riemannian operators, the non-differentiability of their discrete counterparts, and poor parallelisation capabilities have been the main obstacles to the dev

Anna L. Rosen

The stellar initial mass function (IMF) high-mass slope $\alpha$ is routinely measured by fitting single-star models to photometric samples that contain 20-90% unresolved binaries. This practice introduces a systematic negative bias on $\alpha$ that is constant with sample size $N$. Because posterior credible intervals shrink as $1/\sqrt{N}$, at sufficiently

Arnold Mathijssen, Hamed Almohammadi, Lauren Altman, Talia Calazans

Living systems are made of active materials with microscopic components that work together to perform macroscopic biological tasks. The breakdown of these collective functionalities leads to diseases, which, conversely, could be treated by exploiting self-organization in healthcare technologies. Here, we review recent advances in this rapidly growing field o

Yinuo Yang, Zheng Zhang, Ningzhi Tang, Xu Wang

AI-supported writing tools show strong potential for scaffolding students' learning of argumentative writing. Prior work has demonstrated the benefits of AI-supported cognitive scaffolds, such as idea exploration and argument refinement, but how these features function in authentic classroom settings remains underexplored. In this paper, we investigate the c

Konstantinos Dimopoulos, Christian Dioguardi, Ioannis D. Gialamas, Antonio Racioppi

We study quintessential inflation in the framework of metric-affine gravity. It is well known that non-minimal couplings with the Holst invariant can generate a quasi-pole inflationary behaviour resulting in a Starobinsky-like phenomenology. The same quasi-pole behaviour can also be used in order to "flatten" the scalar potential in the Dark Energy era provi

Hugo Parlier, Yandi Wu

This paper is about closed hyperbolic surface amalgams with a focus on the growth of the number of closed geodesics. As in the case of surfaces, we show that topological and volume entropies coincide, but we show stark differences in how they behave according to geometric data with upper and lower bounds on the number of closed geodesics which depend on the

Umar Marikkar, Muhammad Awais, Sara Atito

Computational methods on analyzing Whole Slide Images (WSIs) enable early diagnosis and treatments by supporting pathologists in detection and classification of tumors. However, the extremely high resolution of WSIs makes end-to-end training impractical compared to typical image analysis tasks. To address this, most approaches use pre-trained feature extract

Yara Alakeel, Chatrine Qwaider, Hanan Aldarmaki, Sawsan Alqahtani

This work investigates how effectively large language models (LLMs) and their tokenization schemes represent and generate Arabic root-pattern morphology, probing whether they capture genuine morphological structure or rely on surface memorization. Arabic morphological system provides a rich testbed for analyzing how LLMs handle complex, non-concatenative for

Carlos Rodríguez Cortéz, Moussa Mebarki, Bruno Berini, Dominique Demaille

The recent development of freestanding oxide thin films opens up exciting opportunities for the design of novel heterostructures with enhanced functionalities. Here, we explore the fabrication of membranes consisting of dense arrays of ultrathin CoxNi1-x nanowires epitaxially embedded in a SrTiO3 matrix. Through combined x-ray absorption spectroscopy, x-ray

Yihong Guo, Dongqiangzi Ye, Sijia Chen, Anqi Liu

Autonomous driving requires safe planning, but most learning-based planners lack explicit self-correction ability: once an unsafe action is proposed, there is no mechanism to correct it. Thus, we propose CorrectionPlanner, an autoregressive planner with self-correction that models planning as motion-token generation within a propose, evaluate, and correct lo

Mohsen Sahraei Ardakani, Hong Wan, Rui Song

Mobile Edge Computing (MEC) technology has been introduced to enable could computing at the edge of the network in order to help resource limited mobile devices with time sensitive data processing tasks. In this paradigm, mobile devices can offload their computationally heavy tasks to more efficient nearby MEC servers via wireless communication. Consequently

Michael R. Douglas, Sarah Hoback, Anna Mei, Ron Nissim

A foundational result in constructive quantum field theory is the construction of the free bosonic quantum field theory in four-dimensional Euclidean spacetime and the proof that it satisfies the Glimm-Jaffe axioms, a variant of the Osterwalder-Schrader axioms. We present a formalization of this result in the Lean 4 interactive theorem prover. The project is

F. D. Priestley, P. C. Clark, S. C. O. Glover, S. E. Ragan

The Gao-Solomon relationship between the luminosity of the HCN $J=1-0$ line and the star formation rate (SFR) is observed to remain close to linear over scales ranging from individual star-forming clumps to entire galaxies. This is widely interpreted as the HCN line tracing the reservoir of dense gas directly associated with star formation. However, resolved

Marcell Kegl, Andras Palffy, Csaba Benedek, Dariu M. Gavrila

In this paper, we address extrinsic calibration for camera, lidar, and 4D radar sensors. Accurate extrinsic calibration of radar remains a challenge due to the sparsity of its data. We propose CLRNet, a novel, multi-modal end-to-end deep learning (DL) calibration network capable of addressing joint camera-lidar-radar calibration, or pairwise calibration betw

Ebraheem Farag, Earl P. Bellinger, Philip Mocz, Selim Kalici

We extend the time-dependent convection treatment in \code{MESA} by introducing eddy-viscous damping. This software change brings \code{MESA-TDC} into closer alignment with the radial stellar pulsation framework of \code{MESA-RSP}. We demonstrate that the inclusion of the eddy viscosity in hydrodynamic stellar models remains stable on evolutionary timescales

Manuel Ettengruber, Florian Kühnel

Primordial micro black holes can constitute dark matter if short-distance gravity is modified by extra dimensions or a large number of species and if the memory-burden effect sufficiently suppresses Hawking evaporation. The resulting black holes in the transition regime differ from their four-dimensional Einsteinian counterparts through their mass--radius re

Joshua Davies, Kay Schönwald, Matthias Steinhauser, Daniel Stremmer

We implement the recently-calculated analytic expressions for the next-to-leading order QCD corrections to $gg\to ZH$ in ggxy. This provides a flexible framework for investigating partonic and hadronic cross sections for various top quark mass renormalization schemes. We augment the $Z$ boson with leptonic decays, including spin correlations and off-shell ef

Charlotte Franke, Dorian A. Gangloff

Quantum error correction (QEC) is indispensable for scalable quantum computing, but implementing it with minimal hardware overhead remains a central challenge. Large spin systems with collective degrees of freedom offer a promising route to reducing the control complexity of qubit architectures while retaining a large Hilbert space for fault-tolerant encodin

Jacob Levy, Tyler Westenbroek, Kevin Huang, Fernando Palafox

Robot learning requires adaptation methods that improve reliably from limited, mixed-quality interaction data. This is especially challenging in long-horizon, contact-rich tasks, where end-to-end policy finetuning remains inefficient and brittle. World models offer a compelling alternative: by predicting the outcomes of candidate action sequences, they enabl

Andy J. Goldschmidt, Emilio Peláez Cisneros, Ryan Sitler, Kevin Olsson

We introduce crosstalk-robust gate sets, which are obtained using a novel, scalable optimal control problem exploiting locality. Through the suppression of pairwise quantum crosstalk, the gate sets enable robustness that extends to multi-qubit circuits. The IBM Quantum Platform devices provide a testbed for our gate sets, where we study their efficacy via er

Dhruv Sood, Nilmani Mathur, Vikram Tripathi

The Harrow-Hassidim-Lloyd (HHL) algorithm is a quantum algorithm for solving systems of linear equations that, in principle, offers an exponential improvement in scaling with the system size compared to classical approaches. In this work, we investigate the practical implementation and optimisation of the HHL algorithm with a focus on improving its performan

J. Fernandez, M. Ruhdorfer, J. Serra

We investigate the one-loop renormalization group evolution in four dimensions of the leading operators in the effective field theories of shift-symmetric scalars, photons, and gravitons. We show that certain non-minimal three-point interactions induce a negative running of the corresponding Wilson coefficients, with beta-functions suppressed by the Planck s

Jeein Kim, Aeree Chung, O. Ivy Wong, Junhyun Baek

We present neutral atomic hydrogen (HI) imaging observations of 22 HI-rich ($M_{\rm HI} \gtrsim 10^{9.7} M_\odot$), hard X-ray-selected local Seyferts to explore how cool gas is supplied to active galactic nuclei (AGN) hosts. The sample predominantly resides in group-like, gas-rich environments. About 80% (18/22) of the galaxies have HI-detected neighbors, 6

Matteo Fael, Fabian Lange, Kay Schönwald, Matthias Steinhauser

We calculate the photon-energy spectrum of the inclusive radiative decay $B\to X_s\gamma$, induced by the electromagnetic dipole operator $O_7$, to next-to-next-to-next-to-leading order and consider the complete corrections for light fermions, for the contributions with two closed massive fermion loops, and for the limit of large QCD colour factors $N_{\rm c

Gaia Delucchi, Tullia Sbarrato, Giorgio Calderone, Chiara Righi

Host galaxies of Active Galactic Nuclei give crucial information on the interaction between accreting Supermassive Black Holes and their surroundings, and on their common evolution. Their study in the case of aligned jetted AGN - BL Lacertae objects in particular - is complicated by the non-thermal jet component, whose bright and multi-frequency emission eas

Maximilian Häberle, Vianak Naranjo, Yared Reinarz, Markus Feldt

Despite the emergence of new types of wavefront sensors, the modulated pyramid wavefront sensor remains the workhorse for ELT instrumentation, and is among the options even for advanced high-contrast, high-Strehl instrumentation like PCS and SAXO+. To achieve the required degree of wavefront control, an operation at frequencies of 3kHz, ideally up to 5kHz, i

Stuart Yi-Thomas, David M. Long, Jay D. Sau

Spin-orbit coupling in Bose gases is known to lead to an Ising-symmetry-broken phase where the bosons condense at one of two nonzero momenta. In two dimensions, the finite momentum of the order parameter allows vortex-antivortex pairs that are typically bound in the superfluid phase to freely separate along Ising domain walls. This non-trivial interaction be

Miguel Vanvlasselaer, Sokratis Trifinopoulos, Alexandra P. Klipfel, David I. Kaiser

We investigate how Hawking radiation from low-mass primordial black holes deposits energy into the early-universe plasma and show that the resulting phenomena are hydrodynamic rather than purely diffusive. Combining analytic arguments with relativistic hydrodynamic simulations, we find that the plasma first develops a quasi-steady outflow during the slow eva

Atanu Maity, Andreas Feuerpfeil, Ronny Thomale, Subir Sachdev

Quantum spin liquids are commonly thought to be highly sensitive to lattice geometry, symmetry, and microscopic exchange patterns, leading to a proliferation of seemingly distinct phases across frustrated magnets. Here, we provide a framework that unifies phases that appear distinct from the viewpoint of this intuition. We postulate that the spin-$\tfrac{1}{

Dolly Nambi, Kabir Khanna, Andrew Allocca, Thomas Iadecola

Information-theoretic phase transitions, such as the measurement-induced phase transition (MIPT), characterize the robustness of quantum dynamics to local monitoring and are naturally formulated in terms of trajectories conditioned on typical measurement outcomes, which are naively accessible only through post-selection. Here we implement forced measurements

Xiangyu Cao, Zohar Nussinov

In the theory of decoherence, redundancy is the correlation between a quantum system and fractions of the environment. It underlies the emergence of classical behavior. We show that redundancy can persist despite thermalizing dynamics in the environment. This follows an initial broadcasting interaction that changes the density of a conserved quantity. The mu

Yu-Xin Wang, Anthony J. Brady, Federico Belliardo, Alexey V. Gorshkov

Noise sensing underlies many physical applications including tests of non-classicality, thermometry, verification of correlated phases of quantum matter, and characterization of criticality. While previous works have shown that quantum resources such as entanglement and squeezing can enhance the sensitivity in estimating deterministic signals, less is known

Agnes Valenti, Ina Park, Antoine Georges, Andrew J. Millis

Quantum impurity solvers are the computational bottleneck of quantum embedding approaches to correlated materials, such as dynamical mean-field theory (DMFT). We show that neural networks trained on synthetic, material-agnostic data learn the impurity mapping from hybridization functions and local interactions to Green's functions with quantitative accuracy

Nicholas Geissler, Vladimir Strokov, Christian Kümmerle, Sergey Kushnarev

Next-generation gravitational-wave (GW) detectors, such as the Laser Interferometer Space Antenna (LISA), will observe vast numbers of overlapping signals. Disentangling these signals from instrumental noise and from one another constitutes a significant data analysis challenge. We explore a denoising technique based on embedding time series into Hankel matr

Yoav Zigdon

I provide multiple examples of conformal field theories (CFTs) on the worldsheet that describe string propagation in target space wormholes connecting two disjoint asymptotic manifolds. The worldsheet approach goes beyond the framework of supergravity by incorporating wormholes for which the size of the throat is comparable to the string scale. Typically, st

E. Bellocchi, A. L. Longinotti, Q. Salomé, A. Gil de Paz

The narrow-line Seyfert 1 (NLSy1) galaxy IRAS~17020+4544 is one of the few known sources exhibiting a multi-phase outflow detected in both highly ionized and molecular gas, consistent with AGN feedback operating in an `energy-conserving' regime. We investigate the properties and kinematics of the warm ionized gas using new optical seeing-limited integral-fie

Davis Thuillier, Thomas Scaffidi

Clean two-dimensional Fermi liquids are now known to exhibit an intermediate \emph{tomographic} regime, between ballistic and Navier--Stokes transport, caused by the anomalously slow relaxation of parity-odd multipolar deformations of the Fermi surface. Here we show that this anomaly extends to the dynamical realm. Starting from a microscopic numerical evalu

Zihao Wang, Fangzhou Jiang, Haonan Zheng, Xuejian Shen

The clustering of galaxies encodes key information about the structure and assembly history of their host dark matter (DM) haloes, providing a powerful probe of the origin of extreme high-redshift systems. While halo assembly bias has been extensively studied at low redshift, its behavior in the early Universe remains poorly explored. Using the large-volume,

Valentin Thoss, Abraham Loeb, Andreas Burkert

Observations of the interstellar object 3I/ATLAS have revealed a strong production of gas and dust near perihelion, together with rapid brightening. The outgassing from the nucleus has led to a detectable non-gravitational acceleration. In this work, we combine models of the mass loss rate of water and carbon dioxide to derive the non-gravitational parameter

Sebastian A. R. Ellis, Orion Ning, Nicholas L. Rodd, Jan Schütte-Engel

Black hole superradiance is a powerful probe of ultralight axions. If nature contains a boson with a mass of order $10^{-12}\,$eV, $\textit{mere vacuum fluctuations}$ will lead to its efficient production around spinning stellar mass black holes, forming a gravitational atom that both drains the black hole spin and decays to produce near-monochromatic gravit

Petar Simidzija, Eugene Koskin, Elton Yechao Zhu, Michael Dascal

How can we use a quantum computer to detect the entanglement structure of a quantum state? Bouland et al. (2024) recently provided an algorithm that, given multiple input copies of the state, finds the "hidden cuts"-partitions into fully unentangled qubit registers. Their solution is based on turning cuts into a symmetry which can be detected with a Shor-typ

Adrian E. Bayer, Liam Parker, David Valcin, Shi-Fan Chen

Baryon acoustic oscillations (BAO) underpin the key cosmological results from modern spectroscopic galaxy surveys, but nonlinear gravitational evolution limits the precision achievable with traditional analysis methods. To overcome this, we develop field-level inference for BAO, first reconstructing the initial linear density field and then fitting the BAO s

Iosif Bena, Antoine Bourget, Raphaël Dulac, Dimitrios Toulikas

We reveal the supersymmetric brane configurations that give rise to AdS$_4\times$S$^2\times$S$^2$ supergravity solutions, which are holographic duals to three-dimensional $N=4$ CFTs or to conformal boundaries and domain walls of four-dimensional $N=4$ SYM. We show that these solutions preserve the same Killing spinors as orthogonal D3, D5 and NS5 branes in f

Alberto Castellano, Carmine Montella, Matteo Zatti

We determine the exact functional determinants of charged, massive spin-0 and spin-$\frac12$ particles in $\text{AdS}_2\times \mathbf{S}^2$ backgrounds threaded by constant electric and magnetic fields. This is achieved using Schwinger proper-time formalism, which allows us to derive the full non-perturbative effective action in the 1-loop and constant backg

Soubhik Kumar, Qianshu Lu, Zhong-Zhi Xianyu, Yisong Zhang

Searches for primordial non-Gaussianity (NG) has the potential to not only reveal the physics of cosmic inflation, but also the structure of fundamental interactions at the highest energies. The cosmological collider (CC) physics program exemplifies this possibility and demonstrates how searches for oscillatory NG can lead to mass-spin spectroscopy of extrem

Lianghui Zhu, Yuxin Fang, Bencheng Liao, Shijie Wang

Scaling depth is a key driver for large language models (LLMs). Yet, as LLMs become deeper, they often suffer from signal degradation: informative features formed in shallow layers are gradually diluted by repeated residual updates, making them harder to recover in deeper layers. We introduce mixture-of-depths attention (MoDA), a mechanism that allows each a

Yulin Luo, Hao Chen, Zhuangzhe Wu, Bowen Sui

Vision-Language-Action (VLA) models have recently emerged as a promising paradigm for robotic manipulation, in which reliable action prediction critically depends on accurately interpreting and integrating visual observations conditioned on language instructions. Although recent works have sought to enhance the visual capabilities of VLA models, most approac

Erik Y. Wang, Sumeet R. Motwani, James V. Roggeveen, Eliot Hodges

Can AI make progress on important, unsolved mathematical problems? Large language models are now capable of sophisticated mathematical and scientific reasoning, but whether they can perform novel research is still widely debated and underexplored. We introduce HorizonMath, a benchmark of 113 predominantly unsolved problems spanning eight domains in mathemati

Xincheng Shuai, Ziye Li, Henghui Ding, Dacheng Tao

Generating accurate glyphs for visual text rendering is essential yet challenging. Existing methods typically enhance text rendering by training on a large amount of high-quality scene text images, but the limited coverage of glyph variations and excessive stylization often compromise glyph accuracy, especially for complex or out-of-domain characters. Some m

Lingyu Li, Yan Teng, Yingchun Wang

Existing behavioral alignment techniques for Large Language Models (LLMs) often neglect the discrepancy between surface compliance and internal unaligned representations, leaving LLMs vulnerable to long-tail risks. More crucially, we posit that LLMs possess an inherent state of moral indifference due to compressing distinct moral concepts into uniform probab

Zhenghong Zhou, Xiaohang Zhan, Zhiqin Chen, Soo Ye Kim

Recent video diffusion models have made remarkable strides in visual quality, yet precise, fine-grained control remains a key bottleneck that limits practical customizability for content creation. For AI video creators, three forms of control are crucial: (i) scene composition, (ii) multi-view consistent subject customization, and (iii) camera-pose or object

Yukang Cao, Haozhe Xie, Fangzhou Hong, Long Zhuo

We present HSImul3R, a unified framework for simulation-ready 3D reconstruction of human-scene interactions (HSI) from casual captures, including sparse-view images and monocular videos. Existing methods suffer from a perception-simulation gap: visually plausible reconstructions often violate physical constraints, leading to instability in physics engines an

Aozhe Wang, Yuchen Yan, Nan Zhou, Zhengxi Lu

Reinforcement learning for code generation relies on verifiable rewards from unit test pass rates. Yet high-quality test suites are scarce, existing datasets offer limited coverage, and static rewards fail to adapt as models improve. Recent self-play methods unify code and test generation in a single model, but face a inherent dilemma: white-box access leads

Shixin Wu, Dawei Zhong, Todd A. Brun, Daniel A. Lidar

Code-switching offers a route to universal, fault-tolerant quantum computation by circumventing the limitation implied by the Eastin-Knill theorem against a universal transversal gate set within a single quantum code. Here, we present a fault-tolerant code-switching protocol between two versions of the $[[8, 3, 2]]$ code. One version supports weakly fault-to

Tom A. Rutter, Yuxin Liu, M. Amin Rahimian

Researchers increasingly use data on social and economic networks to study a range of social science questions, but releasing statistics derived from networks can raise significant privacy concerns. We show how to release network connectedness indices that quantify assortative mixing across node attributes under edge-adjacent differential privacy. Standard p

Yi-Ting Lee, Keerthi Kumaran, Bibek Pokharel, Allen Scheie

Realistic simulation of quantum materials is a central goal of quantum computation. Although quantum processors have advanced rapidly in scale and fidelity, it has remained unclear whether pre-fault-tolerant devices can perform quantitatively reliable material simulations. We demonstrate that a superconducting quantum processor operating on up to 50 qubits c

Felix Liedeker, Basil Ell, Philipp Cimiano, Christoph Düsing

Explainability is widely regarded as essential for trustworthy artificial intelligence systems. However, the metrics commonly used to evaluate counterfactual explanations are algorithmic evaluation metrics that are rarely validated against human judgments of explanation quality. This raises the question of whether such metrics meaningfully reflect user perce

Moji Shi, Rajitha de Silva, Hang Yu, Riccardo Polvara

Autonomous exploration in unknown environments typically relies on onboard state estimation for localisation and mapping. Existing exploration methods primarily maximise coverage efficiency, but often overlook that visual-inertial odometry (VIO) performance strongly depends on the availability of robust visual features. As a result, exploration policies can

MiroMind Team, S. Bai, L. Bing, L. Lei

We present MiroThinker-1.7, a new research agent designed for complex long-horizon reasoning tasks. Building on this foundation, we further introduce MiroThinker-H1, which extends the agent with heavy-duty reasoning capabilities for more reliable multi-step problem solving. In particular, MiroThinker-1.7 improves the reliability of each interaction step thro

Jacob Elskamp, Moji Shi, Leonard Bauersfeld, Davide Scaramuzza

Battery-powered multirotor unmanned aerial vehicles (UAVs) can rapidly map unknown environments, but mission performance is often limited by energy rather than geometry alone. Standard exploration policies that optimise for coverage or time can therefore waste energy through manoeuvre-heavy trajectories. In this paper, we address energy-aware autonomous 3D e

Ugur Akcal, Seung Hyun Kim, Mikihisa Yuasa, Hamid Osooli

Spiking neural networks (SNNs) and biologically-inspired learning mechanisms are attractive in mobile robotics, where the size and performance of onboard neural network policies are constrained by power and computational budgets. Existing SNN approaches, such as population coding, reward modulation, and hybrid artificial neural network (ANN)-SNN architecture

Timing Yang, Sicheng He, Hongyi Jing, Jiawei Yang

SAM 3D Body (3DB) achieves state-of-the-art accuracy in monocular 3D human mesh recovery, yet its inference latency of several seconds per image precludes real-time application. We present Fast SAM 3D Body, a training-free acceleration framework that reformulates the 3DB inference pathway to achieve interactive rates. By decoupling serial spatial dependencie

Luca Cocconi, Benoît Mahault, Lorenzo Piro

Smart active agents must allocate finite energetic resources across distinct functions, yet the underlying thermodynamic trade-offs remain poorly understood. Here, we introduce a minimal model of a self-steering particle with an internal polarity-cue sensor coupled to an external environmental field, decomposing its steady-state entropy production rate into

Yidi Wang, Zeyu Ma, Pengcheng Chen, Shiang Fang

Topological insulators host Dirac surface states (SS) protected by time-reversal symmetry. Inter-surface hybridization can gap the SS and give rise to the quantum spin Hall effect in films that are sufficiently thin compared to the SS penetration depth. However, quantifying the SS penetration depth typically requires painstaking synthesis of multiple films w

Jesper Derehag, Carlos Calva, Timmy Ghiurau

Recent conversational memory systems invest heavily in LLM-based structuring at ingestion time and learned retrieval policies at query time. We show that neither is necessary. SmartSearch retrieves from raw, unstructured conversation history using a fully deterministic pipeline: NER-weighted substring matching for recall, rule-based entity discovery for mult

Sašo Grozdanov, Samuel Valach, Mile Vrbica

Bouncing geodesics have been used as valuable probes of black hole singularities. In the dual boundary theory, the presence of bouncing geodesics is encoded in the analytic structure of correlation functions. Thus, when their existence is related to the presence of a black hole singularity, this presents a practical holographic framework to analyse, diagnose

Pengjun Fang, Yingqing He, Yazhou Xing, Qifeng Chen

Existing video-to-audio (V2A) generation methods predominantly rely on text prompts alongside visual information to synthesize audio. However, two critical bottlenecks persist: semantic granularity gaps in training data, such as conflating acoustically distinct sounds under coarse labels, and textual ambiguity in describing micro-acoustic features. These bot

Naoto Tani, Futoshi Futami

We study linear contextual bandits under adversarial corruption and heavy-tailed noise with finite $(1+\epsilon)$-th moments for some $\epsilon \in (0,1]$. Existing work that addresses both adversarial corruption and heavy-tailed noise relies on a finite variance (i.e., finite second-moment) assumption and suffers from computational inefficiency. We propose

Satoshi Tsujimoto, Luc Vinet, Alexei Zhedanov

The most general Ruijsenaars-van Diejen-Takemura Hamiltonians are characterized as Heun operators defined as second order $q$-difference operators with a raising action on elementary rational functions with poles on the Askey-Wilson grid.

Yuwen Du, Rui Ye, Shuo Tang, Xinyu Zhu

Deep search capabilities have become an indispensable competency for frontier Large Language Model (LLM) agents, yet the development of high-performance search agents remains dominated by industrial giants due to a lack of transparent, high-quality training data. This persistent data scarcity has fundamentally hindered the progress of the broader research co

Zachery A. Enderson, Jiyuan Fang, Wei-Chen Wang, Li Xiang

Quantum materials constitute a novel category of substances wherein quantum effects and electron-electron (e-e) interactions give rise to unforeseen phenomena on a macroscopic scale. Of particular interest within the realm of quantum materials are flat bands, which promote heavy conduction electrons and enhance e-e correlation effects. While the engineering

Aurora Maffei, Thomas L. Howarth, Marianna Cafiero, Florence Cameron

The scaling of turbulent premixed flames is typically described by correlations derived for unity-Lewis-number fuels. However, their validity for hydrogen (H$_{2}$) remains uncertain due to the thermodiffusive effects associated with its low Lewis number. In this study, turbulent premixed H$_{2}$ and methane (CH$_{4}$) jet flames are systematically compared

Miao Sun, Alish Kanani, Kaushik Shroff, Umit Ogras

Data movement overheads increase the inference latency of state-of-the-art large language models (LLMs). These models commonly use the bfloat16 (BF16) format for stable training. Floating-point standards allocate eight bits to the exponent, but our profiling reveals that exponent streams exhibit fewer than 3 bits Shannon entropy, indicating high inherent com

Adrian Copetudo, Amon M. Kasper, Tanjung Krisnanda, Gregoire Veyrac

The rich dynamics and large Hilbert space of quantum harmonic oscillators make them natural candidates for hardware-efficient and error-correctable quantum information processing. However, implementing direct entangling operations between oscillators remains an outstanding challenge. Existing strategies typically rely on parametrically activating interaction

Anton Kolonin, Vladimir Krykov

This article presents an overview of approaches to modeling the human psyche in the context of constructing an artificial one. Based on this overview, a concept of cognitive architecture is proposed, in which the psyche is viewed as the operating system of a living or artificial subject, comprising a space of states, including the state of needs that determi

Ludovic Van Waerbeke

The nature of dark matter remains a central problem in cosmology. A compelling possibility is that dark matter is macroscopic, consisting of composite objects formed in the early Universe. We introduce the QCD-AQN framework, a well-motivated scenario in which dark matter is composed of dense aggregates of quark and antiquark matter stabilised by axion domain

Vasiliy A. Es'kin, Egor V. Ivanov

Physics-informed neural networks (PINNs) and neural operators (NOs) for solving the problem of diffraction of Extreme Ultraviolet (EUV) electromagnetic waves from contemporary lithography masks are presented. A novel hybrid Waveguide Neural Operator (WGNO) is introduced, based on a waveguide method with its most computationally expensive components replaced

Junyoung Seo, Hyunwook Choi, Minkyung Kwon, Jinhyeok Choi

What if a world simulation model could render not an imagined environment but a city that actually exists? Prior generative world models synthesize visually plausible yet artificial environments by imagining all content. We present Seoul World Model (SWM), a city-scale world model grounded in the real city of Seoul. SWM anchors autoregressive video generatio

Matej Martinc, Goran Dražič, Anton Kokalj, Katarina Žiberna

Four-dimensional scanning transmission electron microscopy (4D-STEM) provides rich, atomic-scale insights into materials structures. However, extracting specific physical properties - such as polarization directions essential for understanding functional properties of ferroelectrics - remains a significant challenge. In this study, we systematically benchmar

Elizabeth A. Peterson

Quantum sensing of meV-scale scattering and absorption of impinging particles with electrons in solid state detectors is a challenging technological advancement with the potential to enable breakthroughs in quantum information applications and studies of fundamental physics. However, a key obstacle for current sensing schemes is the difficulty in distinguish