March 2025 arXiv papers — page 137

Rui Zhong, Chuang Cheng, Junpeng Xu, Yantong Wei

The evolution from motion capture and teleoperation to robot skill learning has emerged as a hotspot and critical pathway for advancing embodied intelligence. However, existing systems still face a persistent gap in simultaneously achieving four objectives: accurate tracking of full upper limb movements over extended durations (Accuracy), ergonomic adaptatio

Eric Boltersdorf, Thilo vom Hövel, Jeremy Andrew Morín Nenoff, Frank Vewinger

Between the absorption and the emission spectral lineshapes of dense atomic and molecular media, such as dye solutions and alkali-noble buffer gas mixtures at high pressure, in many cases there exists a universal scaling, the Kennard-Stepanov relation, which is a manifestation of detailed balance. This relation plays a crucial role in recent Bose-Einstein co

Giulia Lupi, Seol Ah Park, Martin Ambroz, Resul Ozbilgic

In this paper, we propose a new workflow to analyze macrophage motion during wound healing. These immune cells are attracted to the wound after an injury and they move showing both directional and random motion. Thus, first, we smooth the trajectories and we separate the random from the directional parts of the motion. The smoothing model is based on curve e

Alexander Edström, Paolo Barone, Silvia Picozzi, Massimiliano Stengel

We develop a multiscale approach to magnetoelectric effects, bridging atomistic and continuum models, with all parameters determined from ab initio electronic structure calculations. We show that the parameters of the model are equivalent to the electric field-induced Dzyaloshinski-Moriya interactions. After careful validation, we apply the models to study e

Jeimin Jeon, Youngmin Oh, Junghyup Lee, Donghyeon Baek

N-shot neural architecture search (NAS) exploits a supernet containing all candidate subnets for a given search space. The subnets are typically trained with a static training strategy (e.g., using the same learning rate (LR) scheduler and optimizer for all subnets). This, however, does not consider that individual subnets have distinct characteristics, lead

Vijay Kumar, Martin Siegele-Brown, Parsa Rahimi, Matthew Aylett

We present a novel microfabricated neutral atom source for quantum technologies that can be easily integrated onto microchip devices using well-established MEMS fabrication techniques, and contrast this to conventional off-chip ion loading mechanisms. The heating filament of the device is shown to be as small as 90$\times$90 $\mu$m$^2$. Testing of the $^{171

Youngjin Kwon, Xiao Zhang

As face recognition becomes more widespread in government and commercial services, its potential misuse raises serious concerns about privacy and civil rights. To counteract this threat, various anti-facial recognition techniques have been proposed, which protect privacy by adversarially perturbing face images. Among these, generative makeup-based approaches

Nino Bašić, Patrick W. Fowler, Maxine M. McCarthy, Primož Potočnik

A nut graph is a simple graph whose kernel is spanned by a single full vector (i.e. the adjacency matrix has a single zero eigenvalue and all non-zero kernel eigenvectors have no zero entry). We classify generalisations of nut graphs to nut digraphs: a digraph whose kernel (resp. co-kernel) is spanned by a full vector is dextro-nut (resp. laevo-nut); a bi-nu

Chuqin Geng, Yuhe Jiang, Ziyu Zhao, Haolin Ye

While concept-based explanations improve interpretability over local attributions, they often rely on correlational signals and lack causal validation. We introduce VisionLogic, a novel neural-symbolic framework that produces faithful, hierarchical explanations as global logical rules over causally validated concepts. VisionLogic first learns activation thre

Zixian Liu, Mingtong Zhang, Yunzhu Li

With the rapid advancement of large language models (LLMs) and vision-language models (VLMs), significant progress has been made in developing open-vocabulary robotic manipulation systems. However, many existing approaches overlook the importance of object dynamics, limiting their applicability to more complex, dynamic tasks. In this work, we introduce KUDA,

Vatsal Srivastava

Building upon the concepts and mechanisms used for the development in Moving Points Algorithm, we will now explore how non linear decision boundaries can be developed for classification tasks. First we will look at the classification performance of MPA and some minor developments in the original algorithm. We then discuss the concepts behind using cubic spli

Jing Xu, Franziska Boenisch, Iyiola Emmanuel Olatunji, Adam Dziedzic

Graph Neural Networks (GNNs) have shown remarkable performance in various applications. Recently, graph prompt learning has emerged as a powerful GNN training paradigm, inspired by advances in language and vision foundation models. Here, a GNN is pre-trained on public data and then adapted to sensitive tasks using lightweight graph prompts. However, using pr

Giuseppe D'Onofrio, Anderson Melchor Hernandez

In this work, we consider a multi-population system where the dynamics of each agent evolve according to a system of stochastic differential equations in a general functional setup, determined by the global state of the system. Each agent is associated with a probability measure, that assigns the label accounting for the population to which the agent belongs

Arvid Frydenlund

This work concerns the path-star task, a minimal example of searching over a graph. The graph, $G$, is star-shaped with $D$ arms radiating from a start node, $s$. A language model (LM) is given $G$, $s$, and a target node $t$, which ends one of the arms and is tasked with generating the arm containing $t$. The minimal nature of this task means only a single

Ankit Abhinav, Satyabrata Jana, Abhishek Sahu

In a digraph $D$, an arc $e=(x,y) $ in $D$ is considered transitive if there is a path from $x$ to $y$ in $D- e$. A digraph is transitive-free if it does not contain any transitive arc. In the Transitive-free Vertex Deletion (TVD) problem, the goal is to find at most $k$ vertices $S$ such that $D-S$ has no transitive arcs. In our work, we study a more genera

Vadim Abramkin, George G. Pavlov, Yuriy Shibanov, B. Posselt

Context. The nearby middle-aged gamma-ray pulsar J1741-2054 and its pulsar wind nebula (PWN) have been studied in X-rays, and its bow-shock nebula (BSN) has been investigated in the Balmer lines, but they have never been observed in far ultraviolet (FUV). Aims. To further study the thermal and magnetospheric emission from PSR J1741-2054 and the BSN propertie

Reshma Rastogi, Ankush Bisht, Sanjay Kumar, Suresh Chandra

Support Vector Regression (SVR) and its variants are widely used to handle regression tasks, however, since their solution involves solving an expensive quadratic programming problem, it limits its application, especially when dealing with large datasets. Additionally, SVR uses an epsilon-insensitive loss function which is sensitive to outliers and therefore

Eric C. -Y. Yuan, Yunsheng Liu, Junmin Chen, Peichen Zhong

Given the power of large language and large vision models, it is of profound and fundamental interest to ask if a foundational model based on data and parameter scaling laws and pre-training strategies is possible for learned simulations of chemistry and materials. The scaling of large and diverse datasets and highly expressive architectures for chemical and

Shuo Xie, Tianhao Wang, Sashank Reddi, Sanjiv Kumar

We present a novel unified analysis for a broad class of adaptive optimization algorithms with structured (e.g., layerwise, diagonal, and kronecker-factored) preconditioners for both online regret minimization and offline convex optimization. Our analysis not only provides matching rate to several important structured preconditioned algorithms including diag

Seemant Mishra, Artem Ryabov, Philipp Maass

In driven nonlinear systems, phase locking is an intriguing effect leading to robust stationary states that are stable over extended ranges of control parameters. Recent experiments allow for exploring microscopic mechanisms underlying such phenomena in collective dynamics of micro- and nanoparticles. Here we show that phase-locked dynamics of hardcore-inter

Michael O. Atambo, Korir Kiptiemoi

The zeolitic imidazolate frameworks (ZIF) have emerged as a promising candidate for catalysis, carbon-dioxide (CO$_{2}$) capture and storage as well as flue gas separation due to their tunable porosity and chemical stability. ZIFs consists of transition metals in a tetrahedral coordination with imidazolate linkers, allowing for structural modifications that

Zixuan Liu, Xuyang Wu, Dandan Wang, Jie Lu

This article explores distributed convex optimization with globally-coupled constraints, where the objective function is a general nonsmooth convex function, the constraints include nonlinear inequalities and affine equalities, and the feasible region is possibly unbounded. To address such problems, a unified DUal Consensus Algorithm (DUCA) and its proximal

Robin Schmucker, Steven Moore

High-quality test items are essential for educational assessments, particularly within Item Response Theory (IRT). Traditional validation methods rely on resource-intensive pilot testing to estimate item difficulty and discrimination. More recently, Item-Writing Flaw (IWF) rubrics emerged as a domain-general approach for evaluating test items based on textua

W. Narloch, G. Hajdu, G. Pietrzynski, P. Wielgorski

Type II Cepheids (T2Ceps), alongside RR Lyrae stars, serve as important distance indicators for old population II stars due to their period-luminosity (PL) relations. However, studies of these relations in the Sloan photometric system are rather limited in the literature. Our goal is to calibrate PL relations (and their counterparts in Wesenheit magnitudes)

Sylvain Prolhac

We consider TASEP with a single second class particle and periodic boundary conditions. Using Bethe ansatz, we compute stationary large deviations for the joint statistics of the current of first and second class particles. At large scales, the generating function of the joint cumulants shows an unexpected connection to current fluctuations of TASEP with ope

Quoc-Tien Nguyen, Hong-Hai Nguyen, Van-Thong Huynh

In this study, we present an approach for efficient spatiotemporal feature extraction using MobileNetV4 and a multi-scale 3D MLP-Mixer-based temporal aggregation module. MobileNetV4, with its Universal Inverted Bottleneck (UIB) blocks, serves as the backbone for extracting hierarchical feature representations from input image sequences, ensuring both computa

Alexandra J. Tetarenko, Poshak Gandhi, Devraj Pawar

The most powerful cosmic engines in our universe are fueled by compact objects. These objects accrete large amounts of material and eject matter in the form of jets. Recent groundbreaking discoveries of gravitational waves from merging compact objects and the direct imaging of the black hole shadows with the Event Horizon Telescope represent major steps forw

Zilu Guo, Hongbin Lin, Zhihao Yuan, Chaoda Zheng

3D Multimodal Large Language Models (MLLMs) have recently made substantial advancements. However, their potential remains untapped, primarily due to the limited quantity and suboptimal quality of 3D datasets. Current approaches attempt to transfer knowledge from 2D MLLMs to expand 3D instruction data, but still face modality and domain gaps. To this end, we

Will Boney

The first-order model theory of modules has been studied for decades. More recently, the model theoretic study of nonelementary classes of modules--especially Abstract Elementary Classes of modules--has produced interesting results. This survey aims to discuss these recent results and give an introduction to the framework of Abstract Elementary Classes for m

Luca L. C. Trautmann, Peter A. Wijeratne, Itamar Ronen, Ivor J. A. Simpson

Introducing accelerated reconstruction algorithms into clinical settings requires measures of uncertainty quantification that accurately assess the relevant uncertainty introduced by the reconstruction algorithm. Many currently deployed approaches quantifying uncertainty by focusing on measuring the variability in voxelwise intensity variation. Although thes

Zengrong Lin, Zheng Wang, Tianwen Qian, Pan Mu

Cross-modal retrieval aims to bridge the semantic gap between different modalities, such as visual and textual data, enabling accurate retrieval across them. Despite significant advancements with models like CLIP that align cross-modal representations, a persistent challenge remains: the hubness problem, where a small subset of samples (hubs) dominate as nea

Jun Zhu, Yin Xu, Dazhi He, Haoyang Li

This paper explores the potential of affine frequency division multiplexing (AFDM) to mitigate the multiuser interference (MUI) problem by employing time-domain precoding in extremely-large-scale multiple-input multiple-output (XL-MIMO) systems. In XL-MIMO systems, user mobility significantly improves network capacity and transmission quality. Meanwhile, the

Sebastian Posur

We discuss invariants which are helpful for the computation of the vanishing locus of a finitely presented functor $\mathcal{G}$, i.e., the set of points in the Ziegler spectrum on which $\mathcal{G}$ vanishes. These invariants are: the rank of $\mathcal{G}$, the supports of its co- and contravariant defect, and the class of $\mathcal{G}$ in the Grothendieck

Mohammad Mosaffa, Omid Rafieian, Hema Yoganarasimhan

Political polarization is a significant issue in American politics, influencing public discourse, policy, and consumer behavior. While studies on polarization in news media have extensively focused on verbal content, non-verbal elements, particularly visual content, have received less attention due to the complexity and high dimensionality of image data. Tra

Jun Yu, Yongqi Wang, Lei Wang, Yang Zheng

This paper presents our method for the estimation of valence-arousal (VA) in the 8th Affective Behavior Analysis in-the-Wild (ABAW) competition. Our approach integrates visual and audio information through a multimodal framework. The visual branch uses a pre-trained ResNet model to extract spatial features from facial images. The audio branches employ pre-tr

Zeyue Tian, Zhaoyang Liu, Yizhu Jin, Ruibin Yuan

Audio and music generation based on flexible multimodal control signals is a widely applicable topic, with the following key challenges: 1) a unified multimodal modeling framework, and 2) large-scale, high-quality training data. As such, we propose AudioX, a unified framework for anything-to-audio generation that integrates varied multimodal conditions (i.e.

Enzo Putti-Garcia, Andrii Tykhonov, Andrii Kotenko, Hugo Boutin

The Dark Matter Particle Explorer (DAMPE) is a space-based Cosmic-Ray (CR) observatory with the aim, among others, to study Cosmic-Ray Electrons (CREs) up to 10 TeV. Due to the low CRE rate at multi-TeV energies, we aim to increasing the acceptance by selecting events outside the fiducial volume. The complex topology of non-fiducial events requires the devel

Ana Beatriz Vieira, Maria Valente, Diana Montezuma, Tomé Albuquerque

Quality control of medical images is a critical component of digital pathology, ensuring that diagnostic images meet required standards. A pre-analytical task within this process is the verification of the number of specimen fragments, a process that ensures that the number of fragments on a slide matches the number documented in the macroscopic report. This

Nikolas Liebster, Marius Sparn, Elinor Kath, Jelte Duchene

Driven systems are of fundamental scientific interest, as they can exhibit properties that are radically different from the same system at equilibrium. In certain cases, long-lived states of driven matter can emerge, which exhibit new material properties. In this work, we probe the excitation spectrum of an emergent patterned state in a driven superfluid, fi

Andrew Knight

This paper presents a novel information-theoretic proof demonstrating that the human brain as currently understood cannot function as a classical digital computer. Through systematic quantification of distinguishable conscious states and their historical dependencies, we establish that the minimum information required to specify a conscious state exceeds the

Denis V. Osipov

We consider the group $\mathcal G$ which is the semidirect product of the group of analytic functions with values in ${\mathbb C}^*$ on the circle and the group of analytic diffeomorphisms of the circle that preserve the orientation. Then we construct the central extensions of the group $\mathcal G$ by the group ${\mathbb C}^*$. The first central extension,

Gabriele D'Acunto, Claudio Battiloro

Recent advances in artificial intelligence reveal the limits of purely predictive systems and call for a shift toward causal and collaborative reasoning. Drawing inspiration from the revolution of Grothendieck in mathematics, we introduce the relativity of causal knowledge, which posits structural causal models (SCMs) are inherently imperfect, subjective rep

Fabian Knorr, Philip Salzmann, Peter Thoman, Thomas Fahringer

Parallel programming models can encourage performance portability by moving the responsibility for work assignment and data distribution from the programmer to a runtime system. However, analyzing the resulting implicit memory allocations, coherence operations and their interdependencies can quickly introduce delays into the latency-sensitive execution pipel

Florian Eichin, Yang Janet Liu, Barbara Plank, Michael A. Hedderich

Discourse understanding is essential for many NLP tasks, yet most existing work remains constrained by framework-dependent discourse representations. This work investigates whether large language models (LLMs) capture discourse knowledge that generalizes across languages and frameworks. We address this question along two dimensions: (1) developing a unified

Miguel A. Cardona

In [CMRM24], it was proved that it is relatively consistent that \emph{bounding number} $\mathfrak{b}$ is smaller than the uniformity of $\mathcal{MA}$, where $\mathcal{MA}$ denotes the ideal of the meager-additive sets of $2^{\omega}$. To establish this result, a specific cardinal invariant, which we refer to as $\mathfrak{b}_b^\mathsf{eq}$, was introduced

Uriel Feige, Vadim Grinberg

We consider the problem of fair allocation of $m$ indivisible goods to $n$ agents with either subadditive or XOS valuations, in the arbitrary entitlement case. As fairness notions, we consider the anyprice share (APS) ex-post, and the maximum expectation share (MES) ex-ante. We observe that there are randomized allocations that ex-ante are at least $\frac{1}

Hooman Shahrokhi, Devjeet Raj Roy, Yan Yan, Venera Arnaoudova

We consider the problem of generating valid and small prediction sets by sampling outputs (e.g., software code and natural language text) from a black-box deep generative model for a given input (e.g., textual prompt). The validity of a prediction set is determined by a user-defined binary admissibility function depending on the target application. For examp

Zachary Metzler, Zorawar Wadiasingh

Millisecond pulsars (MSPs) are prolific GeV $\gamma\text{-ray}$ emitters, and nearly 80\% of Fermi-LAT MSPs reside in compact binaries. We demonstrate that the companions in these compact MSPs binaries are also 511 keV annihilation line emitters using {\tt MEGAlib} simulations (a high energy radiation transport software built with {\tt Geant4}) to compute th

Weisong Sun, Yiran Zhang, Jie Zhu, Zhihui Wang

Commenting code is a crucial activity in software development, as it aids in facilitating future maintenance and updates. To enhance the efficiency of writing comments and reduce developers' workload, researchers has proposed various automated code summarization (ACS) techniques to automatically generate comments/summaries for given code units. However, thes

Sahar Admoni, Assaf Hallak, Yftah Ziser, Omer Ben-Porat

Explaining reinforcement learning agents is challenging because policies emerge from complex reward structures and neural representations that are difficult for humans to interpret. Existing approaches often rely on curated demonstrations that expose local behaviors but provide limited insight into an agent's global strategy, leaving users to infer intent fr

Yuhan Wang, Cheng Liu, Daou Zhang, Zihan Zhao

In light of the mounting imperative for public security, the necessity for automated threat detection in high-risk scenarios is becoming increasingly pressing. However, existing methods generally suffer from the problems of uninterpretable inference and biased semantic understanding, which severely limits their reliability in practical deployment. In order t

Jonathan Sejr Pedersen, Andrew Senger

We prove that the Madsen-Tillmann spectrum $MT\theta_n$ splits into the sum of spectra $\Sigma^{-2n}MO\langle n+1 \rangle \oplus \Sigma^{\infty-2n}\mathbb{R} P^\infty_{2n}$ after Postnikov trunctation $\tau_{\leq \ell}$ for $\ell = \lfloor \frac{n}{2} \rfloor - 6$. To accomplish this, we prove that the connecting map in a certain fiber sequence is nullhomoto

Itaï Ben Yaacov, Tomás Ibarlucía

We develop foundational aspects of stability theory in affine logic. On the one hand, we prove appropriate affine versions of many classical results, including definability of types, existence of non-forking extensions, and other fundamental properties of forking calculus. Most notably, stationarity holds over arbitrary sets (in fact, every type is Lascar st

Srivatsav Kunnawalkam Elayavalli, Christopher Schafhauser

We compute the $K_1$-group of ultraproducts of unital, simple $C^*$-algebras with unique trace and strict comparison. As an application, we prove that the reduced free group $C^*$-algebras $C^*_r(F_m)$ and $C^*_r(F_n)$ are elementarily equivalent (i.e., have isomorphic ultrapowers) if and only if $m = n$. This settles in the negative the $C^*$-algebraic anal

Jacob Comeau, Mathieu Bazinet, Pascal Germain, Cem Subakan

Continual learning algorithms aim to learn from a sequence of tasks. In order to avoid catastrophic forgetting, most existing approaches rely on heuristics and do not provide computable learning guarantees. In this paper, we introduce Continual Pick-to-Learn (CoP2L), a method grounded in sample compression theory that retains representative samples for each

Wagner Gomes Rodrigues Junior, Vera Bohomoletz Henriques

Nanoporous capsules have been the subject of intense investigation in the field of drug delivery. One of the essential properties of such particles, which requires characterization, is their structure. Many experimental techniques have been used for this purpose, such as wide-angle neutron or X-ray scattering, or light scattering. Herein, we report theoretic

Mario de Lucio, Jacobo Diaz, Alberto de Castro, Luis E. Romera

Abdominal aortic aneurysms (AAAs) are localized dilatations of the abdominal aorta that can lead to life-threatening rupture if left untreated. AAAs primarily affect older individuals, with high mortality rates following rupture, so early diagnosis and risk assessment are critical. The geometrical characteristics of an AAA, such as its maximum diameter, asym

Xudong Tan, Peng Ye, Chongjun Tu, Jianjian Cao

Multimodal Large Language Models (MLLMs) are becoming increasingly popular, while the high computational cost associated with multimodal data input, particularly from visual tokens, poses a significant challenge. Existing training-based token compression methods improve inference efficiency but require costly retraining, while training-free methods struggle

Jiali Yao, Xinran Deng, Xin Gu, Mengrui Dai

In this paper, we propose spatio-temporal omni-object video grounding, dubbed OmniSTVG, a new STVG task that aims at localizing spatially and temporally all targets mentioned in the textual query from videos. Compared to classic STVG locating only a single target, OmniSTVG enables localization of not only an arbitrary number of text-referred targets but also

John Fernley

The regular tree corresponds to the random regular graph as its local limit. For this reason the famous double phase transition of the contact process on regular tree has been seen to correspond to a phase transition on the large random regular graph, at least at the first critical value. In this article, we find a phase transition on that large finite graph

Michael Schneeberger, Silvia Mastellone, Florian Dörfler

Grid-forming (GFM) converters face significant challenges in limiting current during transient grid events while preserving their grid-forming behavior. This paper offers an elegant solution to the problem with a priori guarantees, presenting a safety filter approach based on Control Barrier Functions (CBFs) to enforce current constraints with minimal deviat

Weihao Xuan, Rui Yang, Heli Qi, Qingcheng Zeng

Existing large language model (LLM) evaluation benchmarks primarily focus on English, while current multilingual tasks lack parallel questions that specifically assess cross-linguistic reasoning abilities. This dual limitation makes it challenging to comprehensively assess LLMs' performance in the multilingual setting. To fill this gap, we introduce MMLU-Pro

Maurizio Grasselli, Luca Melzi, Andrea Signori

In the present work, we develop a comprehensive and rigorous analytical framework for a non-local phase-field model that describes tumour growth dynamics. The model is derived by coupling a non-local Cahn-Hilliard equation with a parabolic reaction-diffusion equation, which accounts for both phase segregation and nutrient diffusion. Previous studies have onl

Hanxu Hu, Jannis Vamvas, Rico Sennrich

LLMs have paved the way for truly simple document-level machine translation, but challenges such as omission errors remain. In this paper, we study a simple method for handling document-level machine translation, by leveraging previous contexts in a multi-turn conversational manner. Specifically, by decomposing documents into segments and iteratively transla

Tom Reichert, Jan Steinheimer, Marcus Bleicher

The violation of isospin symmetry in nucleus-nucleus reactions, as shown in the ratio ${R_K=(K^++K^-)/(K^0+\bar{K}^0)}$ presented by NA61/SHINE, can be understood by introducing results from color-string fragmentation in $e^+e^-$ to nuclear reactions. This novel input allows for a consistent description of the $e^+e^-$ data, proton+proton data and finally nu

Lucas Schorling, Pranav Vaidhyanathan, Jonas Schuff, Miguel J. Carballido

While machine learning holds great promise for quantum technologies, most current methods focus on predicting or controlling a specific quantum system. Meta-learning approaches, however, can adapt to new systems for which little data is available, by leveraging knowledge obtained from previous data associated with similar systems. In this paper, we meta-lear

Alexander V. Osipov

We give a characterization of the $\Delta_1$-property of any Tychonoff space $X$ in terms of the function space $B_1(X)$ of all Baire-one real-valued functions on a space $X$ with the topology of pointwise convergence. We establish that for a Tychonoff space $X$ the $\Delta_1$-property is equivalent to the Choquet property of $B_1(X)$. Also we construct unde

J. K. Wahlstrand, J. E. Sipe

The effect of a constant electric field on two-photon absorption in a semiconductor is calculated using an independent-particle theory. The theoretical framework is an extension of a theory of the one-photon Franz-Keldysh effect [Wahlstrand and Sipe, Phys. Rev. B 82, 075206 (2010)]. The theory includes the effect of the constant field, including field-induce

Shuqi Lu, Xiaohong Ji, Bohang Zhang, Lin Yao

Molecular pretrained representations (MPR) has emerged as a powerful approach for addressing the challenge of limited supervised data in applications such as drug discovery and material design. While early MPR methods relied on 1D sequences and 2D graphs, recent advancements have incorporated 3D conformational information to capture rich atomic interactions.

Evgeniia Vu, Andrei Boiarov, Dmitry Vetrov

Generating co-speech gestures in real time requires both temporal coherence and efficient sampling. We introduce a novel framework for streaming gesture generation that extends Rolling Diffusion models with structured progressive noise scheduling, enabling seamless long-sequence motion synthesis while preserving realism and diversity. Our framework is univer

Jiwei Li, Lingyun Qiu, Zhongjing Wang, Hui Yu

We develop a multiscale framework for estimating sediment concentration in water flow from acoustic wave measurements. At the microscopic scale, the sediment distribution is modeled by a spatially inhomogeneous Poisson cloud, while the quantity of interest is its macroscopic concentration. For the associated random wave model, we derive an effective medium w

Gaurav Kumar Gupta, Pranal Pande, Nirajan Acharya, Aniket Kumar Singh

Large Language Models (LLMs) are revolutionizing medical diagnostics by enhancing both disease classification and clinical decision-making. In this study, we evaluate the performance of two LLM- based diagnostic tools, DeepSeek R1 and O3 Mini, using a structured dataset of symptoms and diagnoses. We assessed their predictive accuracy at both the disease and

Adèle Poudou, Théo Simon, Thomas Montandon, Elsa M. Teixeira

We update constraints on a simple model of self-interacting neutrinos involving a heavy scalar mediator with universal flavor coupling. According to past literature, such a model is allowed by Cosmic Microwave Background (CMB) data, with some CMB and large-scale structure data even favoring a strongly-interacting neutrino (SI$\nu$) scenario over $\Lambda$CDM

Wei Xiao, Shangke Lyu, Zhefei Gong, Renjie Wang

Existing quadrupedal locomotion learning paradigms usually rely on extensive domain randomization to alleviate the sim2real gap and enhance robustness. It trains policies with a wide range of environment parameters and sensor noises to perform reliably under uncertainty. However, since optimal performance under ideal conditions often conflicts with the need

Domenico Pomarico, Alfonso Monaco, Giuseppe Magnifico, Antonio Lacalamita

Grokking is a intriguing phenomenon in machine learning where a neural network, after many training iterations with negligible improvement in generalization, suddenly achieves high accuracy on unseen data. By working in the quantum-inspired machine learning framework based on tensor networks, we numerically prove that grokking phenomenon can be related to an

Ling Xu, Yuliy Baryshnikov, Cynthia Sung

This paper addresses the Dubins path planning problem for vehicles in 3D space. In particular, we consider the problem of computing CSC paths -- paths that consist of a circular arc (C) followed by a straight segment (S) followed by a circular arc (C). These paths are useful for vehicles such as fixed-wing aircraft and underwater submersibles that are subjec

Florian Jarre

This paper considers convex quadratic programs associated with the training of support vector machines (SVM). Exploiting the special structure of the SVM problem a new type of active set method with long cycles and stable rank-one-updates is proposed and tested (CMU: cycling method with updates). The structure of the problem allows for a repeated simple incr

Jiren Sun, Thomas D. Cook

In clinical trials involving both mortality and morbidity, an active treatment can influence the observed risk of the first non-fatal event either directly, through its effect on the underlying non-fatal event process, or indirectly, through its effect on the death process, or both. Discerning the direct effect of treatment on the underlying first non-fatal

Siyin Wang, Zhaoye Fei, Qinyuan Cheng, Shiduo Zhang

Recent advances in large vision-language models (LVLMs) have shown promise for embodied task planning, yet they struggle with fundamental challenges like dependency constraints and efficiency. Existing approaches either solely optimize action selection or leverage world models during inference, overlooking the benefits of learning to model the world as a way

Jacobo Casas-Ramos, Manuel Lama, Manuel Mucientes

In many engineering applications, processes must be followed precisely, making conformance checking between event logs and declarative process models crucial for ensuring adherence to desired behaviors. This is a critical area where Artificial Intelligence (AI) plays a pivotal role in driving effective process improvement. However, computing optimal alignmen

Liyan Luo, Songyan Tian, Lei Wu

A multiscale stochastic-deterministic coupling method is proposed to investigate the complex interactions between turbulent and rarefied gas flows within a unified framework. This method intermittently integrates the general synthetic iterative scheme with the shear stress transport turbulence model into the direct simulation Monte Carlo (DSMC) approach, ena

Cesar Ceballos, Matthias Müller

Brick polytopes constitute a remarkable family of polytopes associated to the spherical subword complexes of Knutson and Miller. They were introduced for finite Coxeter groups by Pilaud and Stump, who used them to produce geometric realizations of generalized associahedra arising from the theory of cluster algebras of finite types. In this paper, we present

Pelle van de Bor, John Brennan, John A. Regan, Jonathan Mackey

The computational expense of solving non-equilibrium chemistry equations in astrophysical simulations poses a significant challenge, particularly in high-resolution, large-scale cosmological models. In this work, we explore the potential of machine learning, specifically Neural Operators, to emulate the Grackle chemistry solver, which is widely used in cosmo

Yingdong Guan, Suguru Yoshida, Jairo Obando-Guevara, Seng Huat Lee

Precise stoichiometry control in single-crystal growth is essential for both technological applications and fundamental research. However, conventional growth methods often face challenges such as non-stoichiometry, compositional gradients, and phase impurities, particularly in non-congruent melting systems. Even in congruent melting systems like Bi2Se3, dev

Cora A. Duggan, Adam Goertz, Adam Polevoy, Mark Gonzales

In this paper, we present Stratified Topological Autonomy for Long-Range Coordination (STALC), a hierarchical planning approach for multi-robot coordination in real-world environments with significant inter-robot spatial and temporal dependencies. At its core, STALC consists of a multi-robot graph-based planner which combines a topological graph with a novel

Shriyank Somvanshi, Anannya Ghosh Tusti, Rohit Chakraborty, Subasish Das

Young motorcyclists, particularly those aged 15 to 24 years old, face a heightened risk of severe crashes due to factors such as speeding, traffic violations, and helmet usage. This study aims to identify key factors influencing crash severity by analyzing 10,726 young motorcyclist crashes in Texas from 2017 to 2022. Two advanced tabular deep learning models

Chao Zhou, Changsheng You, Beixiong Zheng, Xiaodan Shao

In this letter, we propose to deploy rotatable antennas (RAs) at the base station (BS) to enhance both communication and sensing (C&S) performances, by exploiting a new spatial degree-of-freedom (DoF) offered by array rotation. Specifically, we formulate a multi-objective optimization problem to simultaneously maximize the sum-rate of multiple communication

Liming Wu, Wenbing Huang, Rui Jiao, Jianxing Huang

Predicting crystal structures from chemical compositions is a fundamental challenge in materials discovery, complicated by complex 3D geometries that distinguish it from fields like protein folding. Here, we present Diffusion-based Crystal Omni (DAO), a pretrain-finetune framework for crystal structure prediction integrating two Siamese foundation models: a

Abhijeet Sahdev

While utilizing syntactic tools such as parts-of-speech (POS) tagging has helped us understand sentence structures and their distribution across diverse corpora, it is quite complex and poses a challenge in natural language processing (NLP). This study focuses on understanding sentence structure balance - usages of nouns, verbs, determiners, etc - harmonious

Shailesh Lal, Suvajit Majumder, Evgeny Sobko

We introduce a novel machine learning based framework for discovering integrable models. Our approach first employs a synchronized ensemble of neural networks to find high-precision numerical solution to the Yang-Baxter equation within a specified class. Then, using an auxiliary system of algebraic equations, [Q_2, Q_3] = 0, and the numerical value of the Ha

Yifeng Yang, Lin Zhu, Zewen Sun, Hengyu Liu

Out-of-distribution (OOD) detection remains challenging for deep learning models, particularly when test-time OOD samples differ significantly from training outliers. We propose OODD, a novel test-time OOD detection method that dynamically maintains and updates an OOD dictionary without fine-tuning. Our approach leverages a priority queue-based dictionary th

Nicola Gigli

The standard theory of Banach spaces is built upon the notions of vector space, triangle inequality and Cauchy completeness. Here we propose a `hyperbolic' variant of this `elliptic' framework where general linear combinations are replaced by linear combinations with non-negative coefficients, triangle inequality is replaced by reverse triangle inequality an

Tom Maus, Nico Zengeler, Tobias Glasmachers

We present a novel reinforcement learning (RL) environment designed to both optimize industrial sorting systems and study agent behavior in evolving spaces. In simulating material flow within a sorting process our environment follows the idea of a digital twin, with operational parameters like belt speed and occupancy level. To reflect real-world challenges,

Enrique Ruiz Arriola, Pablo Sanchez-Puertas, Christian Weiss

The transverse charge density of the pion is extracted from a dispersive analysis of the $e^+e^- \rightarrow \pi^+\pi^-$ exclusive annihilation data. A logarithmic dispersion relation is used to compute the unknown phase of the timelike pion form factor from the modulus obtained from the annihilation cross section. The method is model-independent and permits

Xunzhi Zheng, Dan Xu

Learning accurate scene reconstruction without pose priors in neural radiance fields is challenging due to inherent geometric ambiguity. Recent development either relies on correspondence priors for regularization or uses off-the-shelf flow estimators to derive analytical poses. However, the potential for jointly learning scene geometry, camera poses, and de

Raphael Holzinger, Nico S. Bassler, Julian Lyne, Fidel G. Jimenez

We present several analytical approaches to the Dicke superradiance problem, which involves determining the time evolution of the density operator for an initially inverted ensemble of $N$ identical two-level systems undergoing collective spontaneous emission. This serves as one of the simplest cases of open quantum system dynamics that allows for a fully an

Ao Chen, Vighnesh Dattatraya Naik, Markus Heyl

Deep neural quantum states have recently achieved remarkable performance in solving challenging quantum many-body problems. While transformer networks appear particularly promising due to their success in computer science, we show that previously reported transformer wave functions haven't so far been capable to utilize their full power. Here, we introduce t

Teresa Conde, Julian Külshammer

This article studies the compatibility of Koenig's notion of an exact Borel subalgebra of a quasi-hereditary or, more generally, standardly stratified algebra with taking idempotent subalgebras or quotients. As an application, we provide bounds for the multiplicities of indecomposable projectives in the principal blocks of BGG category $\mathcal{O}$ having b

Najmeh Miri, Shahrzad Khayatbashi, Jelena Zdravkovic, Amin Jalali

Object-Centric Process Mining (OCPM) enables business process analysis from multiple perspectives. For example, an educational path can be examined from the viewpoints of students, teachers, and groups. This analysis depends on Object-Centric Event Data (OCED), which captures relationships between events and object types, representing different perspectives.

Liang Wen, Yunke Cai, Fenrui Xiao, Xin He

This paper introduces Light-R1, an open-source suite for training long reasoning models using reproducible and cost-effective methodology. Given the proprietary nature of data used in the DeepSeek-R1 series, we develop an alternative approach leveraging exclusively public data and models. Our curriculum training progressively increases data difficulty, combi