May 2025 arXiv papers — page 94

Linxi Zhao, Sofian Zalouk, Christian K. Belardi, Justin Lovelace

Neural language models are black-boxes--both linguistic patterns and factual knowledge are distributed across billions of opaque parameters. This entangled encoding makes it difficult to reliably inspect, verify, or update specific facts. We introduce Limited Memory Language Models (LMLM), a new class of language models that externalizes factual knowledge to

Gabby Litterio, Juan-David Lizarazo-Ferro, Pedro Felzenszwalb, Rashid Zia

We consider the limits of super-resolution using imaging constraints. Due to various theoretical and practical limitations, reconstruction-based methods have been largely restricted to small increases in resolution. In addition, motion-blur is usually seen as a nuisance that impedes super-resolution. We show that by using high-precision motion information, s

Ryo Kamoi, Yusen Zhang, Nan Zhang, Sarkar Snigdha Sarathi Das

Process Reward Models (PRMs) have emerged as a promising approach for improving LLM reasoning capabilities by providing process supervision over reasoning traces. However, existing approaches for constructing PRM training data remain costly and noisy, as they typically rely on human annotation or sampling-based labeling methods that require repeated LLM call

Hongjian Jiang, Anthony W. Lin, Oliver Markgraf, Philipp Rümmer

We present HornStr, the first solver for invariant synthesis for Regular Model Checking (RMC) with the specification provided in the SMT-LIB 2.6 theory of strings. It is well-known that invariant synthesis for RMC subsumes various important verification problems, including safety verification for parameterized systems. To achieve a simple and standardized fi

Ahmed Bouajjani, Wael-Amine Boutglay, Peter Habermehl

We address the problem of verifying automatically procedural programs manipulating parametric-size arrays of integers, encoded as a constrained Horn clauses solving problem. We propose a new algorithmic method for synthesizing loop invariants and procedure pre/post-conditions represented as universally quantified first-order formulas constraining the array e

Chih-Kai Yang, Neo S. Ho, Hung-yi Lee

With advancements in large audio-language models (LALMs), which enhance large language models (LLMs) with auditory capabilities, these models are expected to demonstrate universal proficiency across various auditory tasks. While numerous benchmarks have emerged to assess LALMs' performance, they remain fragmented and lack a structured taxonomy. To bridge thi

Colin P. Lualdi, Spencer J. Johnson, Michael Vayninger, Kristina A. Meier

In classical optical interferometry, loss and background complicate achieving fast nanometer-resolution measurements with illumination at low light levels. Conversely, quantum two-photon interference is unaffected by loss and background, but nanometer-scale resolution is physically difficult to realize. As a solution, we enhance two-photon interference with

Nasim Paykari, Ali Alfatemi, Damian M. Lyons, Mohamed Rahouti

This work explores a novel integration of blockchain methodologies with Wide Area Visual Navigation (WAVN) to address challenges in visual navigation for a heterogeneous team of mobile robots deployed for unstructured applications in agriculture, forestry, etc. Focusing on overcoming challenges such as GPS independence, environmental changes, and computation

Nikolay Fot, Alexander Vinarsky

The article addresses the problem of storing data in extreme environmental conditions with limited computing resources and memory. There is a requirement to create portable, fault-tolerant, modular database management systems (DBMS) that are optimized for use in embedded systems. Existing databases, such as LittleDB, LMDB, and Berkeley DB, are reviewed, and

Mohammad Reza Taesiri, Abhijay Ghildyal, Saman Zadtootaghaj, Nabajeet Barman

With video games now generating the highest revenues in the entertainment industry, optimizing game development workflows has become essential for the sector's sustained growth. Recent advancements in Vision-Language Models (VLMs) offer considerable potential to automate and enhance various aspects of game development, particularly Quality Assurance (QA), wh

Bendong Tan, Tong Su, Yu Weng, Ketian Ye

The increasing integration of renewable energy sources (RESs) and distributed energy resources (DERs) has significantly heightened operational complexity and uncertainty in modern power systems. Concurrently, the widespread deployment of smart meters, phasor measurement units (PMUs) and other sensors has generated vast spatiotemporal data streams, enabling a

Madhura Dutta, Anil Maheshwari, Subhas C. Nandy, Bodhayan Roy

{\em Partial domination problem} is a generalization of the {\em minimum dominating set problem} on graphs. Here, instead of dominating all the nodes, one asks to dominate at least a fraction of the nodes of the given graph by choosing a minimum number of nodes. For any real number $\alpha\in(0,1]$, $\alpha$-partial domination problem can be proved to be NP-

Parth Sarin, Juan Pablo Alperin

A key type of resource needed to address global inequalities in knowledge production and dissemination is a tool that can support journals in understanding how knowledge circulates. The absence of such a tool has resulted in comparatively less information about networks of knowledge sharing in the Global South. In turn, this gap authorizes the exclusion of r

Xiaodan Shao, Rui Zhang, Jihong Park, Tony Q. S. Quek

Six-dimensional movable antenna (6DMA) is an innovative and transformative technology to improve wireless network capacity by adjusting the 3D positions and 3D rotations of antennas/surfaces (sub-arrays) based on the channel spatial distribution. For optimization of the antenna positions and rotations, the acquisition of statistical channel state information

Yuxiang Wei, Yanteng Zhang, Xi Xiao, Tianyang Wang

Decoding visual experiences from fMRI offers a powerful avenue to understand human perception and develop advanced brain-computer interfaces. However, current progress often prioritizes maximizing reconstruction fidelity while overlooking interpretability, an essential aspect for deriving neuroscientific insight. To address this gap, we propose MoRE-Brain, a

Peng Guo, Jaime Park, Frank X. Lee

We aim to explore a more efficient way to simulate few-body dynamics on quantum computers. Instead of mapping the second quantization of the system Hamiltonian to qubit Pauli gates representation via the Jordan-Wigner transform, we propose to use the few-body Hamiltonian matrix under the statevector basis representation which is more economical on the requir

Buket Can Bahadır

In this paper, it is shown that the tameness of the K\"othe space pair $(\lambda^p(A),\lambda^q(B))$ is determined solely by the tameness of the family of quasi-diagonal operators defined between the pair of spaces. We use this tool to fill the gaps in characterization of pairs of power series spaces, adding to the previously established results of Dubinsky,

Wei Zhang, Zhiwei Zhang, Aiyi Liu

The treatment allocation mechanism in a randomized clinical trial can be optimized by maximizing the nonparametric efficiency bound for a specific measure of treatment effect. Optimal treatment allocations which may or may not depend on baseline covariates have been derived for a variety of effect measures focusing on the trial population, the patient popula

Julio H. Toloza, Alfredo Uribe

We obtain sharp asymptotic formulas for the eigenvalues and norming constants of Sturm-Liouville operators associated with the differential expression \[ -\frac{d^2}{dx^2} + x + q(x), \quad x\in [0,\infty), \] together with the boundary condition $\varphi'(0) - b\varphi(0) =0$, $b\in\mathbb{R}$, where \[ q\in \left\{ p\in L^2_{\mathbb{R}}(\mathbb{R}_+,(1+x)^

Mohamed Niged Mabrouk, Daniel Floryan

When groups of inertial swimmers move together, hydrodynamic interactions play a key role in shaping their collective dynamics, including the cohesion of the group. To explore how these interactions influence group cohesion, we develop a three-dimensional, inviscid, far-field model of a swimmer. Focusing on symmetric triangular, diamond, and circular group a

T. Banks

Several papers from the mid to late 1990s suggest that Einstein's equations should be thought of as the hydrodynamic equations of a special class of quantum systems. A classical solution defines subsystems by dividing space-time up into CAUSAL DIAMONDS and Einstein's equations are the hydrodynamics of a system that assigns a density matrix to each diamond wh

Andreas Basse-O'Connor, Mette Skjøtt

The random $k$-SAT problem serves as a model that represents the 'typical' $k$-SAT instances. This model is thought to undergo a phase transition as the clause density changes, and it is believed that the random $k$-SAT problem is primarily difficult to solve near this critical phase. In this paper, we introduce a weak formulation of degrees of freedom for r

Jiahao Qin, Bei Peng, Feng Liu, Guangliang Cheng

Deep learning models frequently encounter feature uncertainty in diverse learning scenarios, significantly impacting their performance and reliability. This challenge is particularly complex in multi-modal scenarios, where models must integrate information from different sources with inherent uncertainties. We propose Dynamic Uncertainty-Aware Learning (DUAL

Alastair Hamilton

We address the treatment of gauge theories within the framework that is formed from combining the machinery of noncommutative symplectic geometry, as introduced by Kontsevich, with Costello's approach to effective gauge field theories within the Batalin-Vilkovisky formalism; discussing the problem of quantization in this context, and identifying the relevant

Adem Limani

We consider threshold phenomenons in the context of weighted $\ell^2$-spaces. Our main result is a summable Baire category version of K\"orner's topological Ivashev-Musatov Theorem, which is proved to be optimal from several aspects.

Adam L. Gross, Sangheon Oh, Minseong Park, T. Patrick Xiao

A programmable linear resistor with a compact footprint would have profound implications for microelectronics, enabling efficient in-sensor analog signal processing and in-memory computing. Non-volatile memory offers a potential solution but suffers from limitations due to the programming mechanisms that confine switching to nanoscale constrictions or field-

Omer Hofman, Jonathan Brokman, Oren Rachmil, Shamik Bose

Agentic AI systems, which build on Large Language Models (LLMs) and interact with tools and memory, have rapidly advanced in capability and scope. Yet, since LLMs have been shown to struggle in multilingual settings, typically resulting in lower performance and reduced safety, agentic systems risk inheriting these limitations. This raises concerns about the

Haoyuan Chen, Emil Constantinescu, Vishwas Rao, Cristiana Stan

The Madden--Julian Oscillation (MJO) is an influential climate phenomenon that plays a vital role in modulating global weather patterns. In spite of the improvement in MJO predictions made by machine learning algorithms, such as neural networks, most of them cannot provide the uncertainty levels in the MJO forecasts directly. To address this problem, we deve

Sarker Md. Sadman, Walter R. L. Lambrecht

LiGa$_5$O$_8$ with a cubic spinel type structure was recently reported to be a ultra-wide-band-gap semiconductor with unintentional p-type conduction. While the origin of p-type doping is still unclear, the fundamental properties of this material are of interest. Here we present a first-principles study of the phonons using density functional perturbation th

Kwang Hak Kim, Mamadou Diagne, Miroslav Krstić

This paper takes a step towards addressing the difficulty of constructing Control Barrier Functions (CBFs) for parallel safety boundaries. A single CBF for both boundaries has been reported to be difficult to validate for safety, and we identify why this challenge is inherent. To overcome this, the proposed method constructs separate CBFs for each boundary.

Morteza Alizadeh, Mehrdad Oveisi, Sonya Falahati, Ghazal Mousavi

Machine learning (ML) models rely heavily on consistent and accurate performance metrics to evaluate and compare their effectiveness. However, existing libraries often suffer from fragmentation, inconsistent implementations, and insufficient data validation protocols, leading to unreliable results. Existing libraries have often been developed independently a

Luc Devroye, Joe R. Hill

We develop uniformly fast random variate generators for the Pearson IV distribution that can be used over the entire range of both shape parameters. Additionally, we derive an efficient algorithm for sampling from the betaized Meixner-Morris density, which is proportional to the product of two generalized hyperbolic secant densities.

Hui Shen, Taiqiang Wu, Qi Han, Yunta Hsieh

Existing benchmarks fail to capture a crucial aspect of intelligence: physical reasoning, the integrated ability to combine domain knowledge, symbolic reasoning, and understanding of real-world constraints. To address this gap, we introduce PhyX: the first large-scale benchmark designed to assess models capacity for physics-grounded reasoning in visual scena

Tony Montes, Fernando Lozano

Recent advancements in Video Question Answering (VideoQA) have introduced LLM-based agents, modular frameworks, and procedural solutions, yielding promising results. These systems use dynamic agents and memory-based mechanisms to break down complex tasks and refine answers. However, significant improvements remain in tracking objects for grounding over time

Nicola Scafetta

The CMIP global climate models (GCMs) assess that nearly 100% of global surface warming observed between 1850-1900 and 2011-2020 is attributable to anthropogenic drivers like greenhouse gas emissions. These models also generate future climate projections based on shared socioeconomic pathways (SSPs), aiding in risk assessment and the development of costly Ne

Awni Altabaa, Omar Montasser, John Lafferty

Learning complex functions that involve multi-step reasoning poses a significant challenge for standard supervised learning from input-output examples. Chain-of-thought (CoT) supervision, which provides intermediate reasoning steps together with the final output, has emerged as a powerful empirical technique, underpinning much of the recent progress in the r

Alejandro M. F Rivas, Eduardo G. Vergini, Leonardo Ermann, Gabriel G. Carlo

The question of how classical thermodynamic laws emerge from the underlying quantum substrate lies at the foundations of physics. Here, we examine the validity of the ideal gas law (IGL) for a single quantum particle confined within a two-dimensional cavity. By interpreting the quantum wave function as a probability density analogous to that of an ideal gas,

Bowen Feng, Zhiting Mei, Julian Ost, Filippo Ghilotti

While autonomous driving (AD) stacks struggle with decision making under partial observability and real-world complexity, human drivers are capable of applying commonsense reasoning to make near-optimal decisions with limited information. Recent work has attempted to leverage finetuned Vision-Language Models (VLMs) for trajectory planning at inference time t

N. Ryde, G. Nandakumar, R. Albarracin, M. Schultheis

The Nuclear Stellar Disc (NSD) is a rotating, disc-like structure in the Galactic Center, believed to have a distinct star formation. However, its formation history and evolutionary links to other structures in the Galactic Center remain uncertain. This study aims to present the first comprehensive chemical census of the NSD by deriving abundance trends for

Silvia Belladitta, Eduardo Bañados, Zhang-Liang Xie, Roberto Decarli

Luminous quasars at $z>4$ provide key insights into the early Universe. Their rarity necessitates wide-field multi-band surveys to efficiently separate them from the main astrophysical contaminants (i.e., ultracool dwarfs). To expand the sample of high-$z$ quasars, we conducted targeted selections using optical, infrared, and radio surveys, complemented by l

Dong Won Lee, Hae Won Park, Cynthia Breazeal, Louis-Philippe Morency

We propose a large language model based reward decomposition framework for aligning dialogue agents using only a single session-level feedback signal. We leverage the reasoning capabilities of a frozen, pretrained large language model (LLM) to infer fine-grained local implicit rewards by decomposing global, session-level feedback. Our first \emph{text-only}

Jenny Ottmann, Frank Breitinger, Felix Freiling

The acquisition of data from main memory or from hard disk storage is usually one of the first steps in a forensic investigation. We revisit the discussion on quality criteria for "forensically sound" acquisition of such storage and propose a new way to capture the intent to acquire an instantaneous snapshot from a single target system. The idea of our defin

Alberto Castellano, Dieter Lüst, Carmine Montella, Matteo Zatti

We compute the supersymmetric entropy of the most general BPS black hole in 4d $\mathcal{N}=2$ supergravity coupled to $n_V$ vector multiplets obtained from Type IIA string theory compactified on a Calabi-Yau threefold at large volume, including the all-genera leading-order $\alpha'$-corrections. These can be equivalently seen as D0-brane quantum effects fro

Xuntao Wu, Yash J. Joshi, Haoxiong Yan, Gustav Andersson

Quantum processors based on superconducting qubits are being scaled to larger qubit numbers, enabling the implementation of small-scale quantum error correction codes. However, catastrophic chip-scale correlated errors have been observed in these processors, attributed to e.g. cosmic ray impacts, which challenge conventional error-correction codes such as th

Aliakbar Nafar, Kristen Brent Venable, Zijun Cui, Parisa Kordjamshidi

In this work, we evaluate the potential of Large Language Models (LLMs) in building Bayesian Networks (BNs) by approximating domain expert priors. LLMs have demonstrated potential as factual knowledge bases; however, their capability to generate probabilistic knowledge about real-world events remains understudied. We explore utilizing the probabilistic knowl

Craig Gidney

Planning the transition to quantum-safe cryptosystems requires understanding the cost of quantum attacks on vulnerable cryptosystems. In Gidney+Eker{\aa} 2019, I co-published an estimate stating that 2048 bit RSA integers could be factored in eight hours by a quantum computer with 20 million noisy qubits. In this paper, I substantially reduce the number of q

Juvenal Domingos Júnior, Augusto Faria, E. Seiti de Oliveira, Erick de Brito

This paper presents BR-TaxQA-R, a novel dataset designed to support question answering with references in the context of Brazilian personal income tax law. The dataset contains 715 questions from the 2024 official Q\&A document published by Brazil's Internal Revenue Service, enriched with statutory norms and administrative rulings from the Conselho Administr

Gavin Stewart

We consider solutions to the Benjamin-Ono equation $$\partial_t u - H \partial_x^2 u = -\partial_x(u^2)$$ that are localized in a reference frame moving to the right with constant speed. We show that any such solution that decays at least like $\langle x\rangle^{-1-\epsilon}$ for some $\epsilon > 0$ in a comoving coordinate frame must in fact decay like $\la

Yuan-Kuei Wu, Juan Azcarreta, Kashyap Patel, Buye Xu

This study presents a deep-learning framework for controlling multichannel acoustic feedback in audio devices. Traditional digital signal processing methods struggle with convergence when dealing with highly correlated noise such as feedback. We introduce a Convolutional Recurrent Network that efficiently combines spatial and temporal processing, significant

Paul K. Faehrmann, Jens Eisert, Richard Kueng

The Hadamard test is naturally suited for the intermediate regime between the current era of noisy quantum devices and complete fault tolerance. Its applications use measurements of the auxiliary qubit to extract information, but disregard the system register completely. Separate advances in classical representations of quantum states via classical shadows a

Zachariah B. Etienne, Thiago Assumpção, Leonardo Rosa Werneck, Samuel D. Tootle

Apparent horizon (AH) finders are essential for characterizing black holes and excising their interiors in numerical relativity (NR) simulations. However, open-source AH finders to date are tightly coupled to individual NR codes. We introduce BHaHAHA, the BlackHoles@Home Apparent Horizon Algorithm, the first open-source, infrastructure-agnostic library for A

Avishai Weizman, Yehuda Ben-Shimol, Itshak Lapidot

ASVspoof challenges are designed to advance the understanding of spoofing speech attacks and encourage the development of robust countermeasure systems. These challenges provide a standardized database for assessing and comparing spoofing-robust automatic speaker verification solutions. The ASVspoof5 challenge introduces a shift in database conditions compar

Maria Luisa Buzzo, Michael Hilker, Anita Zanella, Katja Fahrion

Context. Ultra-diffuse galaxies (UDGs) are an intriguing population of galaxies. Despite their dwarf-like stellar masses and low surface brightness, they have large half-light radii and exhibit a diverse range of globular cluster (GC) populations. Some UDGs host many GCs while others have none, raising questions about the conditions under which star clusters

Alex Kogan

Quantization became a necessary tool for serving ever-increasing Large Language Models (LLMs). RTN (Round-to-Nearest) is perhaps the simplest quantization technique that has been around well before LLMs surged to the forefront of machine learning (ML) research. Yet, it has been largely dismissed by recent and more advanced quantization methods that claim sup

Raditya Weda Bomantara, Ibsal Assi, J. P. F. LeBlanc, Michael Vogl

We present a modification to the bosonic Kitaev chain that, despite being Hermitian, supports both nonHermitian skin effect and nontrivial topological edge modes in its excitation Hamiltonian. We establish an exact mapping between the excitation Hamiltonian of our system and a nonHermitian Su-Schrieffer-Heeger (SSH) model, which allows for a completely analy

Hengyun Zhou, Casey Duckering, Chen Zhao, Dolev Bluvstein

Neutral atom arrays have recently emerged as a promising platform for fault-tolerant quantum computing. Based on these advances, including dynamically-reconfigurable connectivity and fast transversal operations, we present a low-overhead architecture that supports the layout and resource estimation of large-scale fault-tolerant quantum algorithms. Utilizing

Ming Zhang, Cui-Qun Chen, Dao-Xin Yao, Fan Yang

The discovery of superconductivity (SC) with critical temperature $T_c$ above the boiling point of liquid nitrogen in pressurized La$_3$Ni$_2$O$_{7}$ has sparked a surge of exploration of high-$T_c$ superconductors in the Ruddlesden-Popper (RP) phase nickelates. More recently, the RP phase nicklate La$_5$Ni$_3$O$_{11}$, which hosts layered structure with alt

Ameet Gadekar, Suhas Thejaswi

Capacitated fair-range $k$-clustering generalizes classical $k$-clustering by incorporating both capacity constraints and demographic fairness. In this setting, each facility has a capacity limit and may belong to one or more demographic groups. The task is to select $k$ facilities as centers and assign each client to a center such that: ($a$) no center exce

Hengameh Bagherian, Majid Ekhterachian, Stefan Stelzl

We study the implications of precision measurements of light-element abundances, in combination with the Cosmic Microwave Background, for scenarios of physics beyond the Standard Model that generate large inhomogeneities in the baryon-to-photon ratio. We show that precision Big Bang Nucleosynthesis (BBN) places strong constraints on any mechanism that produc

James Gurian, Simon May

We show that the gravothermal collapse of self-interacting dark matter (SIDM) halos can deviate from local thermodynamic equilibrium. As a consequence, the self-similar evolution predicted by the commonly adopted conducting fluid model can be altered or broken. Our results are obtained using a novel, efficient kinetic solver called KiSS-SIDM for tracing the

Mehrad Sahebi, Alice Barthe, Yudai Suzuki, Zoë Holmes

In the quest for quantum advantage, a central question is under what conditions can classical algorithms achieve a performance comparable to quantum algorithms--a concept known as dequantization. Random Fourier features (RFFs) have demonstrated potential for dequantizing certain quantum neural networks (QNNs) applied to regression tasks, but their applicabil

Marta Cocco, Gianluca Grignani, Troels Harmark, Marta Orselli

Binary systems of compact objects in close orbit around a supermassive black hole (SMBH) may form in galactic nuclei, providing a unique environment to probe strong-gravity tidal effects on the binary's dynamics. In this work, we investigate precession resonances arising between the periastron precession frequency of a binary system and its orbital frequenci

Krishan Chand, Gopal-Krishna

The high-luminosity blazar OQ 334 is a leading exponent of the intriguing rare phenomenon of alternating between a flat-spectrum radio quasar (FSRQ) and a BL Lac (BLL) states. Its two optical continuum outbursts observed during the $\sim$ 1.5-year long time span, starting Jan 2018, had been shown to coincide with transition from the FSRQ to BLL state, manife

Christopher Cain, Alexander Van Engelen, Kevin S. Croker, Darby Kramer

Recently, it was pointed out that invoking a large value of the CMB optical depth, $\tau_{\rm CMB} = 0.09$, could help resolve tensions between DESI DR2 BAO data and the CMB. This is larger than the value of $\tau_{\rm CMB} = 0.058$ measured from the Planck low-$\ell$ polarization data. Traditionally, $\tau_{\rm CMB}$ is thought of as a constraint on reioniz

Georgii Paradezhenko, Daniil Rabinovich, Ernesto Campos, Kirill Lakhmanskiy

Variational quantum algorithms have become a standard approach for solving a wide range of problems on near-term quantum computers. Identifying an appropriate ansatz configuration for variational algorithms, however, remains a challenging task, especially when taking into account restrictions imposed by real quantum platforms. This motivated the development

Carl Beadle, Sebastian A. R. Ellis, Jacob M. Leedom, Nicholas L. Rodd

We demonstrate that nuclear magnetic resonance based searches for dark matter (DM) have intrinsic and powerful sensitivity to dark photons and the axion-photon coupling. The reason is conceptually straightforward. An instrument such as CASPEr-Gradient begins with a large sample of nuclear spins polarised in a background magnetic field. In the presence of axi

Z. L. Yang, J. L. Han, D. J. Zhou, W. C. Jing

A stellar common envelope occurs in a binary system when the atmosphere of an evolving star expands to encompass an orbiting companion object. Such systems are predicted to evolve rapidly, ejecting the stellar envelope and leaving the companion in a tighter orbit around a stripped star. We used radio timing to identify a pulsar, PSR J1928+1815, with a spin p

Alberto Rebassa-Mansergas, Enrique Solano, Alex J. Brown, Steven G. Parsons

Binary stars containing a white dwarf and a main-sequence star, WDMS binaries, can be used to study a wide range of aspects of stellar astrophysics. We build a magnitude-limited sample of unresolved WDMS binaries from Gaia DR3 to enlarge these studies. We look for WDMS with available spectra whose location in the Gaia colour-magnitude diagram bridges between

Sopheak Sorn, Yuriy Mokrousov

Altermagnets are an emerging class of unconventional antiferromagnets, characterized by a N\'eel ordering that does not break the translation symmetry of the underlying lattice. Depending on the orientation of the N\'eel vector, the anomalous Hall effect (AHE) may or may not exist. In the so-called pure altermagnets, AHE is forbidden by the magnetic symmetry

A. C. Rubio, K. Breivik, C. Badenes, K. El-Badry

The effectiveness and stability of mass transfer in binaries system are crucial in determining its final product. Rapid binary population synthesis (BPS) codes simplify the complex physics of mass transfer by adopting parameterized prescriptions for the stability of mass transfer, accretion efficiency in stable mass transfer, and the efficiency of common-env

Hong Zhe Chen

The gravitational path integral suggests a striking result: the Hilbert space of closed universes in each superselection sector, a so-called $\alpha$-sector, is one-dimensional. We develop an abstract formalism encapsulating recent proposals that modify the gravitational path integral in the presence of observers and allow larger Hilbert spaces to be associa

Omri Lesser, Sagnik Banerjee, Xuepeng Wang, Jaewon Kim

It is well known that disorder can induce low-energy Andreev bound states in a sign-changing, but fully gapped, superconductor at $\pi-$junctions. Generically, these excitations are localized. Starting from a superconductor with a sign-changing and nodeless order parameter in the clean limit, here we demonstrate a mechanism for increasing the localization le

Aidan Reilly, Alessandro Russo, Philip Schuster, Natalia Toro

We explore the fundamental but untested possibility that the photon is a continuous spin particle (CSP) with a small but non-zero spin Casimir $\rho$. When $\rho\neq 0$, the familiar polarization modes of the photon transform non-trivially under Lorentz boosts, leading to deviations from familiar QED. Surprisingly, these deviations are strongest at low energ

Charles Stahl, Oliver Hart, Alexey Khudorozhkov, Rahul Nandkishore

We introduce a new route to Hilbert space fragmentation in high dimensions leveraging the group-word formalism. We show that taking strongly fragmented models in one dimension and "lifting" to higher dimensions using subsystem symmetries can yield strongly fragmented dynamics in higher dimensions, with subdimensional (e.g., lineonic) excitations. This provid

Valentin Villecroze, Yixin Wang, Gabriel Loaiza-Ganem

The task of quantifying the inherent uncertainty associated with neural network predictions is a key challenge in artificial intelligence. Bayesian neural networks (BNNs) and deep ensembles are among the most prominent approaches to tackle this task. Both approaches produce predictions by computing an expectation of neural network outputs over some distribut

Jake Xuereb

We demonstrate how a single heat exchange between a probe thermal qubit and multi-qubit thermal machine encoding a Boolean function, can determine whether the function is balanced or constant, thus providing a novel thermodynamic solution to the Deutsch-Jozsa problem. We introduce a thermodynamic model of quantum query complexity, showing how qubit thermal m

Raphael Bousso, Sami Kaya

Any gravitating region $a$ in any spacetime gives rise to a generalized entanglement wedge, the hologram $e(a)$. Holograms exhibit properties expected of fundamental operator algebras, such as strong subadditivity, nesting, and no-cloning. But the entanglement wedge EW of an AdS boundary region $B$ with commutant $\bar B$ satisfies an additional condition, c

Noah Braeger, Arun Debray, Markus Dierigl, Jonathan J. Heckman

The U-dualities of maximally supersymmetric supergravity theories lead to celebrated non-perturbative constraints on the structure of quantum gravity. They can also lead to the presence of global symmetries since manifolds equipped with non-trivial duality bundles can carry topological charges captured by non-trivial elements of bordism groups. The recently

Joshua N. Benabou, Katherine Fraser, Mario Reig, Benjamin R. Safdi

Axions, grand unification, and string theory are each compelling extensions of the Standard Model. We show that combining these frameworks imposes strong constraints on the QCD axion mass. Using unitarity arguments and explicit string compactifications - such as those from the Kreuzer-Skarke (KS) type IIB ensemble - we find that the axion mass is favored to

Jakob Moritz

We argue that perturbatively flat vacua (PFVs) introduced in \cite{Demirtas:2019sip} are dual to M-theory compactifications on $G_2$-manifolds, enabling the enumeration of potentially novel $G_2$-manifolds via solutions to Diophantine equations in type IIB flux quanta. Independently, we show that warping corrections to the effective action of type IIB flux v

Fulvio Ferlito, Volker Springel, Christopher T. Davies, Toshiki Kurita

We present a complete forward model of a realistic weak lensing galaxy catalogue based on the 740 Mpc hydrodynamical MillenniumTNG (MTNG) simulation. Starting with a complete particle and cell lightcone covering one octant of the sky with redshift range 0 < $z$ < 1.5, we apply a group and subhalo finder to generate the corresponding galaxy catalogue for a fi

Ryan F. Trainor, Noah R. Lamb, Charles C. Steidel, Yuguang Chen

We present the large-scale spatial Lya profiles of galaxies from the Keck Baryonic Structure Survey (KBSS) at 2<z<3. This work also describes the Lya imaging for the KBSS-Lya survey for the first time. Our sample includes 734 Lya-selected galaxies and 119 continuum-selected galaxies with Lya narrow-band imaging, and we measure the spatial morphology of Lya a

Demetra De Cicco, Gaetano Zazzaro, Stefano Cavuoti, Maurizio Paolillo

Context. A defining characteristic of active galactic nuclei (AGN) that distinguishes them from other astronomical sources is their stochastic variability, which is observable across the entire electromagnetic spectrum. Upcoming optical wide-field surveys, such as the Vera C. Rubin Observatory's Legacy Survey of Space and Time, are set to transform astronomy

Yijie Zheng, Weijie Wu, Qingyun Li, Xuehui Wang

Language-Guided object recognition in remote sensing imagery is crucial for large-scale mapping and automated data annotation. However, existing open-vocabulary and visual grounding methods rely on explicit category cues, limiting their ability to handle complex or implicit queries that require advanced reasoning. To address this issue, we introduce a new su

Zhiyuan Xu, Bohan Li, Huan-ang Gao, Mingju Gao

Generating photorealistic driving videos has seen significant progress recently, but current methods largely focus on ordinary, non-adversarial scenarios. Meanwhile, efforts to generate adversarial driving scenarios often operate on abstract trajectory or BEV representations, falling short of delivering realistic sensor data that can truly stress-test autono

Tong Zheng, Lichang Chen, Simeng Han, R. Thomas McCoy

Human beings naturally utilize multiple reasoning modalities to learn and solve logical problems, i.e., different representational formats such as natural language, code, and symbolic logic. In contrast, most existing LLM-based approaches operate with a single reasoning modality during training, typically natural language. Although some methods explored moda

Penghao Wu, Lewei Lu, Ziwei Liu

Large multimodal models excel in multimodal tasks but face significant computational challenges due to excessive computation on visual tokens. Unlike token reduction methods that focus on token-level redundancy, we identify and study the computation-level redundancy on vision tokens to ensure no information loss. Our key insight is that vision tokens from th

Kuntal Bhandari, Bernard Ducomet, Šarka Nečasová, John Sebastian H. Simon

We consider the Cauchy problem for the barotropic Euler system coupled to a vector Schr\"{o}dinger equation in the whole space. Assuming that the initial density and vector potential are small enough, and that the initial velocity is close to some reference vector field $u_0$ such that the spectrum of $Du_0$ is bounded away from zero, we prove the existence

Federica Arrigoni

Structure from Motion (SfM) refers to the problem of recovering both structure (i.e., 3D coordinates of points in the scene) and motion (i.e., camera matrices) starting from point correspondences in multiple images. It has attracted significant attention over the years, counting practical reconstruction pipelines as well as theoretical results. This paper is

Muquan Yu, Mu Nan, Hossein Adeli, Jacob S. Prince

Understanding functional representations within higher visual cortex is a fundamental question in computational neuroscience. While artificial neural networks pretrained on large-scale datasets exhibit striking representational alignment with human neural responses, learning image-computable models of visual cortex relies on individual-level, large-scale fMR

Satoshi Kosugi

Exemplar-based image colorization aims to colorize a grayscale image using a reference color image, ensuring that reference colors are applied to corresponding input regions based on their semantic similarity. To achieve accurate semantic matching between regions, we leverage the self-attention module of a pre-trained diffusion model, which is trained on a l

Eric J. Michaud, Asher Parker-Sartori, Max Tegmark

We study the problem of creating strong, yet narrow, AI systems. While recent AI progress has been driven by the training of large general-purpose foundation models, the creation of smaller models specialized for narrow domains could be valuable for both efficiency and safety. In this work, we explore two challenges involved in creating such systems, having

Yuqi Zhou, Sunhao Dai, Shuai Wang, Kaiwen Zhou

Recent Graphical User Interface (GUI) agents replicate the R1-Zero paradigm, coupling online Reinforcement Learning (RL) with explicit chain-of-thought reasoning prior to object grounding and thereby achieving substantial performance gains. In this paper, we first conduct extensive analysis experiments of three key components of that training pipeline: input

Ling Yang, Ye Tian, Bowen Li, Xinchen Zhang

We introduce MMaDA, a novel class of multimodal diffusion foundation models designed to achieve superior performance across diverse domains such as textual reasoning, multimodal understanding, and text-to-image generation. The approach is distinguished by three key innovations: (i) MMaDA adopts a unified diffusion architecture with a shared probabilistic for

Xuan Qi, Jiahao Qiu, Xinzhe Juan, Yue Wu

Aligning large language models (LLMs) with human preferences remains a key challenge in AI. Preference-based optimization methods, such as Reinforcement Learning with Human Feedback (RLHF) and Direct Preference Optimization (DPO), rely on human-annotated datasets to improve alignment. In this work, we identify a crucial property of the existing learning meth

Carlos Rodriguez-Pardo, Leonardo Chiani, Emanuele Borgonovo, Massimo Tavoni

We present a neural framework for learning conditional optimal transport (OT) maps between probability distributions. Our approach introduces a conditioning mechanism capable of processing both categorical and continuous conditioning variables simultaneously. At the core of our method lies a hypernetwork that generates transport layer parameters based on the

Patrick Kahardipraja, Reduan Achtibat, Thomas Wiegand, Wojciech Samek

Large language models are able to exploit in-context learning to access external knowledge beyond their training data through retrieval-augmentation. While promising, its inner workings remain unclear. In this work, we shed light on the mechanism of in-context retrieval augmentation for question answering by viewing a prompt as a composition of informational

Sihao Cheng, Jiaxuan Li, Eritas Yang

We report the discovery of a dwarf planet candidate, 2017 OF201, currently located at a distance of 90 au. Its orbit is extremely wide and extends to the inner Oort cloud, with a semi-major axis of 830 au and a perihelion of 45 au, precisely determined from 24 observations over 20 years. Assuming a typical albedo of 0.13, we estimate a diameter about 700 km,

Hwan Chang, Yumin Kim, Yonghyun Jun, Hwanhee Lee

As Large Language Models (LLMs) are increasingly deployed in sensitive domains such as enterprise and government, ensuring that they adhere to user-defined security policies within context is critical-especially with respect to information non-disclosure. While prior LLM studies have focused on general safety and socially sensitive data, large-scale benchmar

Zongzhao Li, Zongyang Ma, Mingze Li, Songyou Li

Multimodal Large Language Models (MLLMs) have demonstrated remarkable capabilities across diverse tasks, yet they lag significantly behind humans in spatial reasoning. We investigate this gap through Transformation-Driven Visual Reasoning (TVR), a challenging task requiring identification of object transformations across images under varying viewpoints. Whil