October 2025 arXiv papers — page 64

Nils Philipp Walter, Chawin Sitawarin, Jamie Hayes, David Stutz

Large Language Models (LLMs) are increasingly deployed in agentic systems that interact with an external environment; this makes them susceptible to prompt injections when dealing with untrusted data. To overcome this limitation, we propose SIC (Soft Instruction Control)-a simple yet effective iterative prompt sanitization loop designed for tool-augmented LL

Pedro Fernando Fernández Espinosa, David Reynoso-Mercado

In this paper some combinatorial and homological tools are used to describe and give an explicit formula for the number of exceptional pairs (exceptional sequences of length two) for some classes of Nakayama Algebras and for the Auslander algebra of a radical square zero algebra of type $\mathbb{A}_n$. In addition, we explore how the number of exceptional pa

Chang Liu, Leona Licht, Jan Rothhardt

We present a theoretical evaluation of radiation dose constraints for extreme ultraviolet (EUV) and soft X-ray microscopy. Our work particularly addresses the long-standing concern regarding strong absorption of EUV radiation in biological specimens. Using an established dose-resolution model, we compare hydrated and dehydrated cellular states and quantify t

A. Besnard, V. Sauvage, S. L. Stever, B. Maffei

As observed on the signal of the Planck-HFI highly sensitive bolometers, the effect of cosmic rays on detectors is a major concern for future similar space missions. Their instruments will have a larger detection surface, increased sensitivity, and more stringent requirements on the suppression of systematic effects. To study the impact of cosmic rays on det

Adetayo Adebimpe, Helmut Neukirchen, Thomas Welsh

Honeypots are decoy systems used for gathering valuable threat intelligence or diverting attackers away from production systems. Maximising attacker engagement is essential to their utility. However research has highlighted that context-awareness, such as the ability to respond to new attack types, systems and attacker agents, is necessary to increase engage

Thomas Welsh, Kristófer Finnsson, Brynjólfur Stefánsson, Helmut Neukirchen

Software supply chains (SSCs) are complex systems composed of dynamic, heterogeneous technical and social components which collectively achieve the production and maintenance of software artefacts. Attacks on SSCs are increasing, yet pervasive vulnerability analysis is challenging due to their complexity. Therefore, threat detection must be targeted, to acco

Maciej Bazarnik, Anika Schlenhoff

Spin filtering and its back-action spin transfer torque (STT) are key ingredients of latest spintronic devices based on magnetic tunnel junctions (MTJs). Resonant tunneling (RT), implemented by design or occurring as parasitic effects, is known to crucially affect macroscopic device performance, but direct experimental access to its individual microscopic pr

S John Sharon Sandeep, Dibyajyoti Sahu, Suhas Gangadharaiah

We examine excitation suppression in the transverse-field Ising model (TFIM), where finite-time drive across a quantum critical point is assisted by the presence of a time-dependent coupling parameter. While conventional counterdiabatic protocols are designed to eliminate excitations, they often require complex many-body terms that are difficult to realize e

Jie Zhang, Xiaohong Li, Mengke Zhang, Ruitao Feng

Blockchain-based Attribute-Based Access Control (BC-ABAC) offers a decentralized paradigm for secure data governance but faces two inherent challenges: the transparency of blockchain ledgers threatens user privacy by enabling reidentification attacks through attribute analysis, while the computational complexity of policy matching clashes with blockchain&#39

M. J. Adriaans, J. P. Hoogenboom, A. Mohammadi-Gheidari

Monochromators are an essential component in electron microscopy and spectroscopy for enhancing the spatial and energy resolution. However, its adoption in scanning electron microscopes remains limited because of its high cost and operational complexity. Through a thin-deflector analysis of an electrostatic homogeneous-field deflector, the extreme sensitivit

Francesco Lisi, Pierdomenico Duttilo, Marina Bertolini

Integrating energy islands into the European electricity market is a key challenge for the energy transition. This study investigates the impact of the Sorgente-Rizziconi interconnector on electricity price volatility in Sicily. Before its commissioning on 28 May 2016, the Sicilian electricity market zone was poorly interconnected with the Italian mainland.

Bingru Huang, Falai Chen

In this article, we study the dimension of the spline space of di-degree $(d,d)$ with the highest order of smoothness over a hierarchical T-mesh $\mathscr T$ using the smoothing cofactor-conformality method. Firstly, we obtain a dimensional formula for the conformality vector space over a tensor product T-connected component. Then, we prove that the dimensio

Johannes Tölle, Marios-Petros Kitsaras, Pierre-François Loos

The Bethe-Salpeter equation (BSE) formalism, combined with the $GW$ approximation for ionization energies and electron affinities, is emerging as an efficient and accurate method for predicting optical excitations in molecules. In this letter, we present the first derivation and implementation of fully analytic nuclear gradients for the BSE@$G_0W_0$ method.

Davood Marripour, Jahanfar Abouie

Time quasicrystals (TQCs) represent a compelling extension of the concept of time crystals (TCs). While TCs break discrete time-translation symmetry by exhibiting a periodic response at a subharmonic of the driving frequency, TQCs display a more complex temporal order. They respond at multiple incommensurate frequencies, values that are not integer multiples

Oleg Mushkarov, Nikolai Nikolov

We discuss the optimization problem for minimizing the $(n-1)$-volume of the intersection of a convex cone $K$ in $\Bbb R^n$ with a hyperplane through a given point, first considered in \cite{We}. We give a geometric characterization of the stationary hyperplanes for this problem when $K$ is a hyperangle which partially answers a question posed in \cite{We}.

Adrien Brochier, Lukas Woike

Modular functors are traditionally defined as systems of projective representations of mapping class groups of surfaces that are compatible with gluing. They can formally be described as modular algebras over central extensions of the modular surface operad, with the values of the algebra lying in a suitable symmetric monoidal $(2,1)$-category $\mathcal{S}$

Faraz Zargari, Hossein Nekouyan, Lyndon Hallett, Bo Sun

We study the online multi-class selection problem with group fairness guarantees, where limited resources must be allocated to sequentially arriving agents. Our work addresses two key limitations in the existing literature. First, we introduce a novel lossless rounding scheme that ensures the integral algorithm achieves the same expected performance as any f

Hamid Abdollahi, James Fowler, Carlos Uribe, Arman Rahmim

Predictive dosimetry is central to enabling personalized radiopharmaceutical therapy (RPT), particularly in prostate specific membrane antigen (PSMA) targeted theranostics. In this work, we develop a three layer computational framework that integrates physiologically based pharmacokinetic (PBPK) modeling with machine learning (ML) to predict both physical (A

Li An, Yujian Liu, Yepeng Liu, Yuheng Bu

Watermarking has emerged as a promising solution for tracing and authenticating text generated by large language models (LLMs). A common approach to LLM watermarking is to construct a green/red token list and assign higher or lower generation probabilities to the corresponding tokens, respectively. However, most existing watermarking algorithms rely on heuri

Daniel M. Steinberg, Asiri Wijesinghe, Rafael Oliveira, Piotr Koniusz

We introduce active generation of Pareto sets (A-GPS), a new framework for online discrete black-box multi-objective optimization (MOO). A-GPS learns a generative model of the Pareto set that supports a-posteriori conditioning on user preferences. The method employs a class probability estimator (CPE) to predict non-dominance relations and to condition the g

Rebecca G. Hart, Wanjiku A. Makumi, Rushikesh Kamalapurkar, Warren E. Dixon

Deep neural networks (DNNs) are powerful black-box function approximators which have been shown to yield improved performance compared to traditional neural network (NN) architectures. However, black-box algorithms do not incorporate known physics of the system and can yield results which are physically implausible. Physics-informed neural networks (PINNs) h

Emmanuel Gnandi

We investigate the global topology of 3-dimensional Hessian manifolds. We prove that any compact, orientable 3-dimensional Hessian manifold is either a Hantzsche-Wendt manifold or admits the structure of a K\"ahler mapping torus. We analyze the parity of Betti numbers for compact, orientable 3-dimensional Hessian manifolds, with special focus on those of Kos

Atoosa Chegini, Hamid Kazemi, Garrett Souza, Maria Safi

Reasoning has become a central paradigm for large language models (LLMs), consistently boosting accuracy across diverse benchmarks. Yet its suitability for precision-sensitive tasks remains unclear. We present the first systematic study of reasoning for classification tasks under strict low false positive rate (FPR) regimes. Our analysis covers two tasks--sa

Jiabo Shi, Dimitrios Pezaros, Yehia Elkhatib

The global scarcity of GPUs necessitates more sophisticated strategies for Deep Learning jobs in shared cluster environments. Accurate estimation of how much GPU memory a job will require is fundamental to enabling advanced scheduling and GPU sharing, which helps prevent out-of-memory (OOM) errors and resource underutilization. However, existing estimation m

Robyn Wyrick

As artificial intelligence (AI) advances toward superhuman capabilities, aligning these systems with human values becomes increasingly critical. Current alignment strategies rely largely on externally specified constraints that may prove insufficient against future super-intelligent AI capable of circumventing top-down controls. This research investigates wh

Ziyang Liu, Ning Hao, Yue Selena Niu, Han Xiao

Testing for the presence of autocorrelation is a fundamental problem in time series analysis. Classical methods such as the Box-Pierce test rely on the assumption of stationarity, necessitating the removal of non-stationary components such as trends or shifts in the mean prior to application. However, this is not always practical, particularly when the mean

Zlatan Ajanović, Ravi Prakash, Leandro de Souza Rosa, Jens Kober

Learning from demonstration has proved itself useful for teaching robots complex skills with high sample efficiency. However, teaching long-horizon tasks with multiple skills is challenging as deviations tend to accumulate, the distributional shift becomes more evident, and human teachers become fatigued over time, thereby increasing the likelihood of failur

Ali Khosravi Kazazi, Zhenlong Li, M. Naser Lessani, Guido Cervone

The complexity of SQL and the spatial semantics of PostGIS create barriers for non-experts working with spatial data. Although large language models can translate natural language into SQL, spatial Text-to-SQL is more error-prone than general Text-to-SQL because it must resolve geographic intent, schema ambiguity, geometry-bearing tables and columns, spatial

Kunal Shankar, Ninad Gaikwad, Anamika Dubey

Achieving the flexibility from house heating, cooling, and ventilation systems (HVAC) has the potential to enable large-scale demand response by aggregating HVAC load adjustments across many homes. This demand response strategy helps distribution grid to flexibly ramp-up or ramp-down local load demand so that it can optimally match the bulk power system gene

Benjamin Lange

When should we defer to AI outputs over human expert judgment? Drawing on recent work in social epistemology, I motivate the idea that some AI systems qualify as Artificial Epistemic Authorities (AEAs) due to their demonstrated reliability and epistemic superiority. I then introduce AI Preemptionism, the view that AEA outputs should replace rather than suppl

Valmir P Moraes Filho, Daniela Martin, Jasmine R. Kobayashi, Connor O'Brien

The solar wind is a dynamic plasma outflow that shapes heliospheric conditions and drives space weather. Identifying its large-scale phenomena is crucial, yet the increasing volume of high-cadence Parker Solar Probe (PSP) observations poses challenges for scalable, interpretable analysis. We present a pipeline combining symbolic compression, density-based cl

Zijian Zhang, Rong Wang, Shiyang Li, Yuebo Luo

Developing efficient CUDA kernels is increasingly critical for AI applications such as large-scale LLM training. However, manual kernel design is both costly and time-consuming, motivating automatic approaches that leverage LLMs for code generation. Existing methods for automatic kernel generation, however, often produce low-efficiency kernels, incur high co

Yamil Cahuana Medrano, Hervé Dutrieux, Joseph Karpie, Kostas Orginos

The extraction of parton distribution functions (PDFs) from experimental or lattice QCD data is an ill-posed inverse problem, where regularization strongly impacts both systematic uncertainties and the reliability of the results. We study a framework based on Gaussian Process Regression (GPR) to reconstruct PDFs from lattice QCD matrix elements. Within a Bay

Mohammad Mahdi Danesh Pajouh, Sara Saeedi

Meningiomas represent the most prevalent form of primary brain tumors, comprising nearly one-third of all diagnosed cases. Accurate delineation of these tumors from MRI scans is crucial for guiding treatment strategies, yet remains a challenging and time-consuming task in clinical practice. Recent developments in deep learning have accelerated progress in au

Christopher Jerrett, Elliot Anshelevich

We study the problem of selecting a representative committee of $k$ agents from a collection of $n$ agents in a common metric space. This problem is related to choosing $k$ facilities in facility location and $k$-median problems. However, unlike in more traditional facility location where each agent only cares about the closest selected facility, in the sett

Gereon Elvers, Gilad Landau, Oiwi Parker Jones

Non-invasive brain-computer interfaces (BCIs) are beginning to benefit from large, public benchmarks. However, current benchmarks target relatively simple, foundational tasks like Speech Detection and Phoneme Classification, while application-ready results on tasks like Brain-to-Text remain elusive. We propose Keyword Spotting (KWS) as a practically applicab

L. C. N. Santos

In recent years, there has been a growing interest in the study of regular black holes, driven by the search for singularity-free geometries. This research has revealed intriguing similarities between the regularization mechanisms used in black hole models and those employed in quantum field theory, such as the introduction of exponential suppression or ener

P. Varghese, S. Raman, M. Guran, L. Reyes

PIP-II is a superconducting linac that is in the initial acceleration chain for the Fermilab accelerator complex. The RF system consists of a warm front-end with an RFQ and buncher cavities along with 25 superconducting cryo-modules comprised of cavities with five different acceleration \(\beta\). The LLRF system for the linac has to provide field and resona

Wagner Cortes, Eduardo N. Marcos

In this article, we introduce the concept of partial actions of a group $G$ on quivers and demonstrate that for any given partial action of G on a quiver $\Gamma$, there exists another quiver, $\Gamma'$ with a full $G$-action. This is an enveloping action of the partial action of $G$ on $\Gamma$. We also introduce partial actions of groups on algebras by sub

Barkavi Sundararajan, Somayajulu Sripada, Ehud Reiter

A major concern when deploying LLMs in accuracy-critical domains such as sports reporting is that the generated text may not faithfully reflect the input data. We quantify how input structure affects hallucinations and other factual errors in LLM-generated summaries of NBA play-by-play data, across three formats: row-structured, JSON and unstructured. We man

Willem Diepeveen, Melanie Weber

We develop a theory of iso-Riemannian optimization for problems constrained to learned data manifolds, a setting in which classical Riemannian optimization - and Riemannian gradient descent in particular - can be poorly suited. That is, favorable Euclidean properties of an objective need not translate into geodesic convexity or L-smoothness, and the Riemanni

Michel Orsi, Veeraj Shah, Mahesh Padmanabhan, Thomas Curwen

We experimentally investigate the rheology of dark chocolate pastes in both industrially relevant pre-refined form and simplified model systems. Steady and oscillatory shear experiments reveal yielding, pronounced shear-thinning, and stress-dependent hysteresis governed by solid loading. Fitting the viscosity data with the Maron-Pierce model provides stress-

Qinghua Lu, Dehai Zhao, Yue Liu, Hao Zhang

The emergence of foundation models (FMs) has enabled the development of highly capable and autonomous agents, unlocking new application opportunities across a wide range of domains. Evaluating the architecture of agents is particularly important as the architectural decisions significantly impact the quality attributes of agents given their unique characteri

En-Jui Chang

Introducing controlled overlap among a few repetition blocks yields a fourfold asymptotic rate improvement while preserving an average stabilizer weight of \(4\). Substituting the overlapped outer layer with an LDPC code further produces a family of constructions with asymptotic rate \(2/d\). We also describe a constant-excitation variant that suppresses col

Kyle Matkovic, Patrick Russell, Andrew Palmer, Eric Helgemo

Electrons levitating above the surface of solid neon have recently emerged as a promising platform for high-quality qubits. The morphology and uniformity of the neon growth in these systems is crucial for qubit performance in a scalable architecture. Here we report on the controlled growth and characterization of thin solid neon films using multiplexed super

Amna Al-Araimi, Yue Zheng, Haiming Liu

This paper discusses the issue regarding Non-verbal Autism Spectrum Disorder. It has been observed that this mental disorder is listed in major parts of the world including the US, UK, and India. To mitigate this type of disorder, a wide range of smartphones, computers, and artificial intelligence technologies have been used. This technology has helped the p

Zhe Fei, Mehmet Yigit Turali, Shreyas Rajesh, Xinyang Dai

Harmonizing medication data across Electronic Health Record (EHR) systems is a persistent barrier to monitoring medications for opioid use disorder (MOUD). In heterogeneous EHR systems, key prescription attributes are scattered across differently formatted fields and freetext notes. We present a practical framework that customizes open source large language

Sai Haneesh Allu, Jishnu Jaykumar P, Ninad Khargonkar, Tyler Summers

We introduce a novel system for human-to-robot trajectory transfer that enables robots to manipulate objects by learning from human demonstration videos. The system consists of four modules. The first module is a data collection module that is designed to collect human demonstration videos from the point of view of a robot using an AR headset. The second mod

Fred Zimmerman

Xynapse Traces is an experimental publishing imprint created via a fusion of human and algorithmic methods using a configuration-driven architecture and a multi-model AI integration framework. The system achieved a remarkable 90% reduction in time-to-market (from a typical 6-12 months to just 2-4 weeks), with 80% cost reduction compared to traditional imprin

Ahmed Saad Al-Karsani, Maryam Khanbaghi, Aleksandar Zečević

The transition towards clean energy and the introduction of Distributed Energy Resources (DERs) are giving rise to the emergence of Microgrids (MGs) and Networks of MGs (NMGs). MGs and NMGs can operate autonomously in islanded mode. However, they face challenges in terms of secondary level frequency and voltage regulation, due to the variable nature of Renew

Jonathan Gold, Tristan Freiberg, Haruna Isah, Shirin Shahabi

The integration of machine learning (ML) systems into critical industries such as healthcare, finance, and cybersecurity has transformed decision-making processes, but it also brings new challenges around trust, security, and accountability. As AI systems become more ubiquitous, ensuring the transparency and correctness of AI-driven decisions is crucial, esp

Qingsong Xu, Jonathan L Bamber, Nils Thuerey, Niklas Boers

Accurate long-term forecasting of spatiotemporal dynamics remains a fundamental challenge across scientific and engineering domains. Existing machine learning methods often neglect governing physical laws and fail to quantify inherent uncertainties in spatiotemporal predictions. To address these challenges, we introduce a physics-consistent neural operator (

Jasmine R. Kobayashi, Daniela Martin, Valmir P Moraes Filho, Connor O'Brien

Labeling or classifying time series is a persistent challenge in the physical sciences, where expert annotations are scarce, costly, and often inconsistent. Yet robust labeling is essential to enable machine learning models for understanding, prediction, and forecasting. We present the \textit{Clustering and Indexation Pipeline with Human Evaluation for Reco

Xiaoxin Ye, Chengkai Huang, Hongtao Huang, Lina Yao

Users increasingly interact with content across multiple domains, resulting in sequential behaviors marked by frequent and complex transitions. While Cross-Domain Sequential Recommendation (CDSR) models two-domain interactions, Multi-Domain Sequential Recommendation (MDSR) introduces significantly more domain transitions, compounded by challenges such as dom

Konstantinos Christopher Tsiolis, Alireza Mousavi-Hosseini, Murat A. Erdogdu

To understand feature learning dynamics in neural networks, recent theoretical works have focused on gradient-based learning of Gaussian single-index models, where the label is a nonlinear function of a latent one-dimensional projection of the input. While the sample complexity of online SGD is determined by the information exponent of the link function, rec

Wanhao Yu, Zheng Wang, Shuteng Niu, Sen Lin

Zeroth-order (ZO) optimization has gained attention as a memory-efficient alternative to first-order (FO) methods, particularly in settings where gradient computation is expensive or even impractical. Beyond its memory efficiency, in this work, we investigate ZO optimization for continual learning (CL) as a novel approach to address the plasticity-stability-

Panayiotis Kousoulas, Rahul Sharma, Y. B. Guo

Fatigue behaviors of metal components by laser fusion suffer from scattering due to random geometrical defects (e.g., porosity, lack of fusion). Monitoring fatigue crack initiation and growth is critical, especially for laser-fused components with significant inherent fatigue scattering. Conventional statistics-based curve-fitting fatigue models have difficu

Yuhong Luo, Austin Hoag, Xintong Wang, Philip S. Thomas

Representation learning is increasingly applied to generate representations that generalize well across multiple downstream tasks. Ensuring fairness guarantees in representation learning is crucial to prevent unfairness toward specific demographic groups in downstream tasks. In this work, we formally introduce the task of learning representations that achiev

Peter Howard, Alim Sukhtayev

For many applications, critical information about system dynamics is encoded in associated eigenvalue problems that can be posed as linear Hamiltonian systems with suitable boundary conditions. Motivated by examples from hydrodynamics, quantum mechanics, and magnetohydrodynamics (MHD), we develop a general framework for analyzing a broad class of linear Hami

Ari Frummer, Helin Wang, Tianyu Cao, Adi Arbel

Source separation is a crucial pre-processing step for various speech processing tasks, such as automatic speech recognition (ASR). Traditionally, the evaluation metrics for speech separation rely on the matched reference audios and corresponding transcriptions to assess audio quality and intelligibility. However, they cannot be used to evaluate real-world m

Stéphane Mathis

From the Earth's atmosphere and oceans to stellar radiation zones, inertia-gravity waves, which are called gravito-inertial waves (hereafter GIWs) in Astrophysics, are transporting momentum and mixing matter when they are damped through heat and viscous diffusions and when they break. Their short-time scale dynamics is governed by the buoyancy force and the

Fatima Badmos, Emma Murphy, Michael Ward, Damon Berry

Existing research and physical activity guidelines highlight the benefits of outdoor physical activities for ageing populations. There is potential for technology to facilitate outdoor activity through Physical Web infrastructure. We proposed that embedding Physical Web applications that are engaging and interactive in public open spaces as part of interacti

William Lauga, James Rowbottom, Alexander Denker, Željko Kereta

We present a framework for solving a broad class of ill-posed inverse problems governed by partial differential equations (PDEs), where the target coefficients of the forward operator are recovered through an iterative regularization scheme that alternates between FEM-based inversion and learned graph neural regularization. The forward problem is numerically

Ludovic Salomon, Daniel Dörfler, Andreas Löhne

MOCVXPY is an open-source Python library for convex vector optimization. It is built on top of CVXPY, a domain-specific language for single-objective convex optimization. MOCVXPY enables practitioners to describe their convex vector optimization problem in an intuitive algebraic language, that closely follows the mathematical formulation. This work presents

Pedro D. González Pérez, Miguel Robredo Buces

In this paper we describe the multiplier ideals and jumping numbers associated with an irreducible germ of quasi-ordinary hypersurface $(D, 0) \subset (\mathbb{C}^{d+1}, 0)$ by using a toroidal embedded resolution. The approach is motivated by Howald's description of the multiplier ideals of monomial ideals. We show that the multiplier ideals of $D$ can be e

Yiliu Li, Esteban Rojas-Gatjens, Yinjie Guo, Birui Yang

Van der Waals (vdW) heterostructures of two-dimensional (2D) materials have become a rich playground for the exploration of correlated quantum phases, and recent studies have begun to probe their non-equilibrium dynamics under femtosecond laser excitation. In a time-resolved experiment, optical excitation of the multilayer structure can lead not only to rich

Samuel Lewis-Lim, Xingwei Tan, Zhixue Zhao, Nikolaos Aletras

Chain-of-thought (CoT) prompting is a common technique for improving the reasoning abilities of large language models (LLMs). However, extended reasoning is often unnecessary and substantially increases token usage. As such, a key question becomes how to optimally allocate compute to when reasoning is actually needed. We study this through confidence-gated C

Yiliu Li, Esteban Rojas-Gatjens, Yinjie Guo, Birui Yang

Photoexcitation has been utilized to control quantum matter and to uncover metastable phases far from equilibrium. Among demonstrations to date, the most common is the photo-induced transition from correlated insulators to metallic states; however, the reverse process without initial orders has not been observed. Here, we show ultrafast metal-to-insulator tr

Nguyen Linh Bao Nguyen, Alsharif Abuadbba, Kristen Moore, Tingmin Wu

The rapid advancement of generative models has enabled the creation of increasingly stealthy synthetic voices, commonly referred to as audio deepfakes. A recent technique, FOICE [USENIX'24], demonstrates a particularly alarming capability: generating a victim's voice from a single facial image, without requiring any voice sample. By exploiting correlations b

Enshu Liu, Qian Chen, Xuefei Ning, Shengen Yan

Image Auto-regressive (AR) models have emerged as a powerful paradigm of visual generative models. Despite their promising performance, they suffer from slow generation speed due to the large number of sampling steps required. Although Distilled Decoding 1 (DD1) was recently proposed to enable few-step sampling for image AR models, it still incurs significan

Alberto-Mario Castillo, Gabriel Villalobos

The movement of the infective form of the T. Cruzi parasite within the human blood is not completely understood. Video microscopy observations confirm forward motility of the protozoa and relate it to the deformation of the body, nonetheless there are open questions relating the deformation of the protozoan with its motion in blood, for which a computational

Jiaqi Hu, Hongli Xu, Junwen Huang, Peter KT Yu

Accurate 6D pose estimation is essential for robotic manipulation in industrial environments. Existing pipelines typically rely on off-the-shelf object detectors followed by cropping and pose refinement, but their performance degrades under challenging conditions such as clutter, poor lighting, and complex backgrounds, making detection the critical bottlenec

Shamim Yazdani, Akansha Singh, Nripsuta Saxena, Zichong Wang

In recent years, deep learning based generative models, particularly Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Diffusion Models (DMs), have been instrumental in in generating diverse, high-quality content across various domains, such as image and video synthesis. This capability has led to widespread adoption of these model

Ilmar Gahramanov, Sinan Ulaş Öztürk, Uveys Turhan

In this work, we develop new Bailey pairs for the pentagon identity satisfied by the tetrahedron index, expressible in terms of $q$-series. Since the tetrahedron index underlies topological invariants of 3-manifolds and related knots, our construction may offer a new framework to deriving knot invariants through the Bailey chain.

Martin Andersson, Anubhab Chowdhury, Erik G. Larsson

We present a framework for joint amplification and phase shift optimization of the repeater gain in dynamic time-division duplex (TDD) repeater-assisted massive MIMO networks. Repeaters, being active scatterers with amplification and phase shift, enhance the received signal strengths for users. However, they inevitably also amplify undesired noise and interf

James Ghawaly, Andrew Nicholson, Catherine Schuman, Dalton Diez

We present a general framework for training spiking neural networks (SNNs) to perform binary classification on multivariate time series, with a focus on step-wise prediction and high precision at low false alarm rates. The approach uses the Evolutionary Optimization of Neuromorphic Systems (EONS) algorithm to evolve sparse, stateful SNNs by jointly optimizin

Xiaohong Chen, Min Seong Kim, Sokbae Lee, Myung Hwan Seo

We propose SLIM (Stochastic Learning and Inference in overidentified Models), a scalable stochastic approximation framework for nonlinear GMM. SLIM forms iterative updates from independent mini-batches of moments and their derivatives, producing unbiased directions that ensure almost-sure convergence. It requires neither a consistent initial estimator nor gl

Ignacio Boero, Ignacio Hounie, Alejandro Ribeiro

Despite the non-convexity of most modern machine learning parameterizations, Lagrangian duality has become a popular tool for addressing constrained learning problems. We revisit Augmented Lagrangian methods, which aim to mitigate the duality gap in non-convex settings while requiring only minimal modifications, and have remained comparably unexplored in con

Jesimon Barreto, Carlos Caetano, André Araujo, William Robson Schwartz

Foundation models have advanced computer vision by enabling strong performance across diverse tasks through large-scale pretraining and supervised fine-tuning. However, they may underperform in domains with distribution shifts and scarce labels, where supervised fine-tuning may be infeasible. While continued self-supervised learning for model adaptation is c

Maria Ciudad Alañón, Daniel Centeno, Andrew Watford, Elie Wolfe

The certification of intrinsic randomness is foundational to quantum information theory and central in many practical applications thereof, such as in the generation of unquestionably random numbers and in cryptographic protocols. Device-independent randomness certification based on violations of Bell inequalities has been thoroughly investigated within the

Mark Phillip Matovic

This work applies Generative Flow Networks (GFlowNets) to three graph optimization problems: the Traveling Salesperson Problem, Minimum Spanning Tree, and Shortest Path. GFlowNets are generative models that learn to sample solutions proportionally to a reward function. The models are trained using the Trajectory Balance loss to build solutions sequentially,

Flora Roumpani, Joel Dearden, Alan Wilson

The two-tier Lowry model brings dynamic simulations of population and employment directly into the planning process. By linking regional modelling with neighbourhood design, the framework enables planners to explore how alternative planning scenarios may evolve over time. The upper tier captures regional flows of people, jobs, and services, while the lower t

Ward Struyve

Bose et al. and Marletto and Vedral proposed an experiment to test whether gravity can induce entanglement between massive systems, arguing that the capacity to do so would imply the quantum nature of gravity. In this work, a class of semi-classical models is examined that treat gravity classically, through some potential in the Schr\"odinger equation, and i

Ian Anderson, Jack Kramer, Tzu-Hsuan Hsu, Yinan Wang

Frequency combs consist of a spectrum of evenly spaced spectral lines. Optical frequency combs enable technologies ranging from timing, LiDAR, and ultra-stable signal sources. Microwave frequency combs are analogous to optical frequency combs, but often leverage electronic nonlinearity for comb generation. Generating microwave frequency combs using piezoelec

A. J. Drake, S. G. Djorgovski, M. J. Graham, D. Stern

The Catalina Real-time Transient Survey (CRTS) carried out a public survey for optical transients between 2007 and 2019, discovering more than 16,000 transient candidates. Here we present the spectra and highlight the results of the spectroscopic follow-up of CRTS extragalactic transients. As expected, we find that the bulk of these transients are normal sup

Anh Pham, Mihir Thalanki, Michael Sun, Aditya Chaloo

Large language models often lose previously aligned safety behaviors when fine-tuned on benign data, a phenomenon known as catastrophic forgetting. Prior work shows that adding random safety examples can mitigate this effect, but it remains unclear which examples are most effective. We propose a behavior-aware sampling framework that selects safety examples

Shuang Geng, Wenli Zhang, Jiaheng Xie, Rui Wang

Social media user-generated content (UGC) provides real-time, self-reported indicators of mental health conditions such as depression, offering a valuable source for predictive analytics. While prior studies integrate medical knowledge to improve prediction accuracy, they overlook the opportunity to simultaneously expand such knowledge through predictive pro

Kalevi Mursula

Sunspots are the standard measure of solar magnetic activity, which are also used to estimate solar spectral irradiance over centennial time scales. However, because of the lack of homogeneous, century-long spectral measurements, the long-term relation of sunspots and spectral irradiance has not been independently validated. Here we aim to study the relation

Thomas D. DeWitt, Timothy J. Garrett

Conceptual and theoretical models describing the dynamics of the atmosphere often assume a hierarchy of dynamic regimes, each operating over some limited range of spatial scales. The largest scales are presumed to be governed by quasi-two-dimensional geostrophic turbulence, mesoscale dynamics by gravity waves, and the smallest scales by 3D isotropic turbulen

V. A. Chizhikov, V. E. Dmitrienko

A new theoretical approach has been developed to describe the elastic properties of cubic blue phases of cholesteric liquid crystals (LCs). Blue phases are three-dimensional periodic chiral liquids with local anisotropy of the average orientation of molecules, and due to their periodicity, they have lattice elastic moduli characteristic of ordinary crystalli

David Lagziel, Ehud Lehrer

We examine information structures in settings with privately informed agents and an informationally constrained mediator who supplies additional public signals. Our focus is on characterizing the set of posteriors that the mediator can induce. To this end, we employ a graph-theoretic framework: states are represented as vertices, information sets correspond

Avinash Patil

Large Language Models (LLMs) increasingly produce natural language explanations alongside their predictions, yet it remains unclear whether these explanations reference predictive cues present in the input text. In this work, we present an empirical study of how LLM-generated explanations align with predictive lexical evidence from an external model in text

Chao Wang, Zhizhao Wen, Ruoxin Zhang, Puyang Xu

In response to the increasingly critical demand for accurate prediction of GPU memory resources in deep learning tasks, this paper deeply analyzes the current research status and innovatively proposes a deep learning model that integrates bidirectional gated recurrent units (BiGRU) to optimize the Transformer architecture, aiming to improve the accuracy of m

Xi Zhang, Xiaolin Wu, Jiamang Wang, Weisi Lin

Large Language Models (LLMs) have demonstrated remarkable capabilities but typically require extensive computational resources and memory for inference. Post-training quantization (PTQ) can effectively reduce these demands by storing weights in lower bit-width formats. However, standard uniform quantization often leads to notable performance degradation, par

Joris Edelmann, Giovanni P. Galdi, Mher M. Karakouzian, Thomas Richter

We perform analytical and numerical analyses of the propulsion of a rigid body in a viscous fluid subjected to a periodic force with zero average over a period. This general formulation specifically addresses the significant case, where propulsion is generated by the oscillation of a mass located in an internal cavity of the body. We provide a rigorous proof

Stefan Abi-Karam, Rishov Sarkar, Suhail Basalama, Jason Cong

Dataflow hardware designs enable efficient FPGA implementations via high-level synthesis (HLS), but correctly sizing first-in-first-out (FIFO) channel buffers remains challenging. FIFO sizes are user-defined and balance latency and area-undersized FIFOs cause stalls and potential deadlocks, while oversized ones waste memory. Determining optimal sizes is non-

Matthew E. Caplan, Nevin T. Smith, Dany Yaacoub, Roberto F. Serrano

We present calculations of diffusion coefficients in grain boundaries in Yukawa crystals for astrophysics. Our methods follow from our recent work calculating diffusion coefficients in perfect body-centered cubic crystals. These diffusion coefficients show only a weak dependence on the crystal orientations at the grain boundary and are consistent with those

Aël Quélennec, Pavlo Mozharovskyi, Van-Tam Nguyen, Enzo Tartaglione

On-device neural network training faces critical memory constraints that limit the adaptation of pre-trained models to downstream tasks. We present MeDyate, a theoretically-grounded framework for memory-constrained dynamic subnetwork adaptation. Our approach introduces two key innovations: LaRa (Layer Ranking), an improved layer importance metric that enable

Ayoub El Hanchi, Murat Erdogdu, Chris Maddison

What property of the data distribution determines the excess risk of principal component analysis? In this paper, we provide a precise answer to this question. We establish a central limit theorem for the error of the principal subspace estimated by PCA, and derive the asymptotic distribution of its excess risk under the reconstruction loss. We obtain a non-

Wence Jiang, Hui Li, Nahuel Andrés, Lina Hadid

Turbulence is fundamental to energy transfer across scales in space and astrophysical plasmas. Bow shock interactions have long been hypothesized to significantly modify turbulence in planetary environments, yet the quantification of such effects and their parametric dependencies remain largely unaddressed. Using in situ long-term high-time resolution measur