May 2025 arXiv papers — page 120

Masamichi Miyaji, Soichiro Mori, Kazumi Okuyama

We extend the notion of chord number in the strict large $N$ double-scaled Sachdev-Ye-Kitaev (DSSYK) model to the corresponding finite $N$ ETH matrix model. The chord number in the strict large $N$ DSSYK model is known to correspond to the discrete length of the Einstein-Rosen bridge in the gravity dual, which reduces to the renormalized geodesic length in J

Krzysztof Frączek, Adam Kanigowski, Corinna Ulcigrai

We consider smooth flows preserving a smooth invariant measure, or, equivalently, locally Hamiltonian flows on compact orientable surfaces, and show that almost every such locally Hamiltonian flow with only simple saddles has singular spectrum. Furthermore, we prove that for almost every pair of such flows, the elements of the pair are spectrally disjoint. M

Christoph Jürgen Hemmer, Daniel Durstewitz

Complex, temporally evolving phenomena, from climate to brain activity, are governed by dynamical systems (DS). DS reconstruction (DSR) seeks to infer generative surrogate models of these from observed data, reproducing their long-term behavior. Existing DSR approaches require purpose-training for any new system observed, lacking the zero-shot and in-context

Pengcheng Pan, Yonekura Shogo, Yasuo Kuniyoshi

Inspired by foveal vision, hard attention models promise interpretability and parameter economy. However, existing models like the Recurrent Model of Visual Attention (RAM) and Deep Recurrent Attention Model (DRAM) failed to model the hierarchy of human vision system, that compromise on the visual exploration dynamics. As a result, they tend to produce atten

Omer Shavit, Hervé Suaudeau, Carine Julien, Hodaya Klimovsky

The back-focal plane (BFP) of a high-numerical aperture objective contains the fluoro-phore radiation pattern, which encodes information about the axial fluorophore position, molecular orientation and the local refractive index of the embedding medium. BFP image acquisition and analysis are common to conoscopy, k-space imaging, supercritical-angle fluorescen

Giacomo Greco

We adopt a Gamma and Malliavin Calculi point of view in order to generalize Score-based diffusion Generative Models (SGMs) to an infinite-dimensional abstract Hilbertian setting. Particularly, we define the forward noising process using Dirichlet forms associated to the Cameron-Martin space of Gaussian measures and Wiener chaoses; whereas by relying on an ab

Juntian Zhu, Miguel de Carvalho, Zhouwang Yang, Fengxiang He

An AI agent might surprisingly find she has reached an unknown state which she has never been aware of -- an unknown unknown. We mathematically ground this scenario in reinforcement learning: an agent, after taking an action calculated from value functions $Q$ and $V$ defined on the {\it {aware domain}}, reaches a state out of the domain. To enable the agent

Elena Sammarco

In the moduli space $\mathcal{C}$ of complex cubic hypersurfaces $X\subset\mathbb{P}^5$, we study the condition that $X$ admits a net of polar quadrics whose discriminant locus is a $10$-nodal irreducible plane sextic curve. Our main result is that such a condition defines an irreducible divisor in $\mathcal{C}$ which is not of Noether-Lefschetz type.

Philipp Scholl, Alexander Dietrich, Sebastian Wolf, Jinoh Lee

Accurately modeling the friction torque in robotic joints has long been challenging due to the request for a robust mathematical description. Traditional model-based approaches are often labor-intensive, requiring extensive experiments and expert knowledge, and they are difficult to adapt to new scenarios and dependencies. On the other hand, data-driven meth

Matteo Buttarazzi, Claudia Ceci

This paper develops a continuous-time filtering framework for estimating a hazard rate subject to an unobservable change-point. This framework naturally arises in both financial and insurance applications, where the default intensity of a firm or the mortality rate of an individual may experience a sudden jump at an unobservable time, representing, for insta

Christopher J. Fewster, Harkan J. Kirk-Karakaya

Quantum backflow is a surprising phenomenon in which a quantum particle, moving in one dimension and with a state of rightwards momentum, can exhibit a net probability transfer to the left-hand half-line over a finite time interval. We generalise the setting of quantum backflow to allow for $M$ disjoint time intervals, considering the sum of probability diff

Natalia A. Boitsova, Anna A. Abelit, Daniil D. Stupin

Nowadays electrical impedance spectroscopy (EIS) has become an advanced experimental technique with a wide range of applications: from simple passive circuits diagnostics to semiconductor high-end device development and breakthrough technologies in bio-sensing. Although hardware for EIS today is well developed, the EIS analysis software is mainly custom, old

Jeffrey R. Forshaw, Simon Plätzer, Fernando Torre González

We study differential intra-jet radiation patterns in jet production at full colour. We present a systematic study of several QCD $2\to 2$ processes and also multi-jet production from a colourless initial state. We examine how subleading colour corrections are distributed differentially in phase space and find that mere normalization effects due to subleadin

Jianfeng Xu

This paper addresses the current lack of a unified formal framework in machine learning theory, as well as the absence of robust theoretical foundations for interpretability and ethical safety assurance. We first construct a formal information model, employing sets of well-formed formulas (WFFs) to explicitly define the ontological states and carrier mapping

Zhengrui Ma, Yang Feng, Chenze Shao, Fandong Meng

We introduce SLED, an alternative approach to speech language modeling by encoding speech waveforms into sequences of continuous latent representations and modeling them autoregressively using an energy distance objective. The energy distance offers an analytical measure of the distributional gap by contrasting simulated and target samples, enabling efficien

Matteo Merler, Nicola Dainese, Minttu Alakuijala, Giovanni Bonetta

Integrating Large Language Models with symbolic planners is a promising direction for obtaining verifiable and grounded plans, with recent works extending this idea to visual domains using Vision-Language Models (VLMs). However, an open-source benchmark for comparing these approaches under matched conditions is missing, due to a lack of visual benchmarks tha

Shuo Cao, Ao Wang, Zheyong Fan, Hua Bao

Metals play a crucial role in heat management in electronic devices, such as integrated circuits, making it vital to understand heat transport in elementary metals and alloys. In this work, we systematically study phonon thermal transport in 16 metals using the efficient homogeneous nonequilibrium molecular dynamics (HNEMD) method and the recently developed

Amanda Stricklan, Tim Waters, James Klimchuk

To model the temperature evolution of optically thin astrophysical environments at MHD scales, radiative and collisional cooling rates are typically either pre-tabulated or fit into a functional form and then input into MHD codes as a radiative loss function. Thermal balance requires estimates of the analogous heating rates, which are harder to calculate, an

Maria Gabriela Boada, Andrea Delgado, Jose Morales Escalante

The transmon qubit, essential to quantum computation, exhibits disordered dynamics under strong parametric drives critical to its control. We present a combined theoretical and numerical study of stability regions in circuit QED using Floquet theory, focusing on the appearance of Arnold tongues that distinguish stable from unstable regimes. Starting from sim

Zihao Cheng, Hongru Wang, Zeming Liu, Yuhang Guo

While integrating external tools into large language models (LLMs) enhances their ability to access real-time information and domain-specific services, existing approaches focus narrowly on functional tool selection following user instructions, overlooking the context-aware personalization in tool selection. This oversight leads to suboptimal user satisfacti

Siming Sun, Kai Zhang, Xuejun Jiang, Wenchao Meng

The emerging paradigm of leveraging pretrained large language models (LLMs) for time series forecasting has predominantly employed linguistic-temporal modality alignment strategies through token-level or layer-wise feature mapping. However, these approaches fundamentally neglect a critical insight: the core competency of LLMs resides not merely in processing

Alp Eren Sari, Paolo Favaro

We propose FlowCut, a simple and capable method for unsupervised video instance segmentation consisting of a three-stage framework to construct a high-quality video dataset with pseudo labels. To our knowledge, our work is the first attempt to curate a video dataset with pseudo-labels for unsupervised video instance segmentation. In the first stage, we gener

V. S. D. S. Mahesh Akavarapu, Hrishikesh Terdalkar, Pramit Bhattacharyya, Shubhangi Agarwal

Large Language Models (LLMs) have demonstrated remarkable generalization capabilities across diverse tasks and languages. In this study, we focus on natural language understanding in three classical languages -- Sanskrit, Ancient Greek and Latin -- to investigate the factors affecting cross-lingual zero-shot generalization. First, we explore named entity rec

Sara Monsurrò, Carmen Perugia, Federica Raimondi

We analyse the effect of a Signorini-type interface condition on the asymptotic behaviour, as {\epsilon} tends to zero, of a problem posed in an open bounded cylinder of {R^N}, {N\geq 2}, divided in two connected components by an imperfect rough surface. The Signorini-type condition is expressed by means of two complementary equalities involving the jump of

Yixuan Xu, Antoni-Joan Solergibert i Llaquet, Antoine Bosselut, Imanol Schlag

Large language models are known to memorize parts of their training data, posing risk of copyright violations. To systematically examine this risk, we pretrain language models (1B/3B/8B) from scratch on 83B tokens, mixing web-scale data with public domain books used to simulate copyrighted content at controlled frequencies at lengths at least ten times longe

Sara Alosaime, Arshad Jhumka

Federated Learning (FL) is a privacy-preserving machine learning technique that allows decentralized collaborative model training across a set of distributed clients, by avoiding raw data exchange. A fundamental component of FL is the selection of a subset of clients in each round for model training by a central server. Current selection strategies are myopi

Marc Kegel, Lisa Piccirillo

Distinct knots K, K' can sometimes share a common p/q-framed Dehn surgery. A folk conjecture held that for a fixed pair of knots, this can occur for at most one value of p/q. We disprove this conjecture by constructing pairs of distinct knots K,K' that have common Dehn surgeries for four distinct slopes. We also construct non-isotopic Legendrian knots K,K' t

J. Thainá-Batista, R. Cid Fernandes, R. M. González Delgado, J. E. Rodríguez-Martín

Spatially resolved maps of stellar populations and nebular emission are key tools for understanding the physical properties and evolutionary stages of galaxies. We aim to characterize the spatially resolved stellar population and emission line properties of galaxies in the M101 group using Javalambre Photometric Local Universe Survey (J-PLUS) data. The datac

Ollie Thakar

We study the moduli space of solutions to the Seiberg-Witten equations with $N$ spinors on a compact Riemann surface. These moduli spaces arise in a program to define a new enumerative invariant of 3-manifolds. They are also of independent interest in the geometry of algebraic curves, as they parameterize generalized divisors in Brill-Noether theory for high

Sara Alosaime, Arshad Jhumka

Federated Learning (FL) enables collaborative model training while preserving privacy by allowing clients to share model updates instead of raw data. Pervasive computing environments (e.g., for Human Activity Recognition, HAR), which we focus on in this paper, are characterized by resource-constrained end devices, streaming sensor data and intermittent clien

Tokuhiro Eto, Harald Garcke, Robert Nürnberg

In this study, we propose a parametric finite element method for a degenerate multi-phase Stefan problem with triple junctions. This model describes the energy-driven motion of a surface cluster whose distributional solution was studied by Garcke and Sturzenhecker. We approximate the weak formulation of this sharp interface model by an unfitted finite elemen

Iris Kaplan, Or Ordentlich

Recent work have shown that the quantization for matrix multiplication problem can be optimally solved by quantizing each column in each matrix using a nested lattice code, and then multiplying the de-quantized matrices. It was further demonstrated that when product codes of sub-dimension $d$ and rate $R$ are used, the de-quantization and inner product opera

Saiei-Jaeyeong Matsubara-Heo

Given a family of varieties, the Euler discriminant locus distinguishes points where Euler characteristic differs from its generic value. We introduce a hypergeometric system associated with a flat family of very affine locally complete intersection varieties. It is proven that the Euler discriminant locus is its singular locus and is purely one-codimensiona

Qi-Hong Cai, Xue-Hao Yu, Ma-Cheng Yang, Ao-Xiang Liu

This study investigates the emergence of macroscopic classical behavior from quantum foundations via the entropic Leggett--Garg inequality. We introduce a geometric framework for deriving entropic Leggett--Garg inequalities with higher-order temporal correlations and demonstrate their advantages over conventional formulations. Numerical analyses show that en

Kyosuke S. Sato, Yutaka Komiyama, Sakurako Okamoto, Masafumi Yagi

We derive the star formation history (SFH) and chemical evolution history (CEH) of the Ursa Minor (UMi) dwarf spheroidal galaxy (dSph). We detect two distinct stellar populations that exist over 6 times half-light radius from its center. The results are obtained by applying a newly developed algorithm to the deep and wide-field photometric dataset taken with

Diogo Landau, Jorge Barbosa, Nishant Saurabh

Online Data Intensive applications (e.g. message brokers, ML inference and databases) are core components of the modern internet, providing critical functionalities to connecting services. The load variability and interference they experience are generally the main causes of Quality of Service (QoS) degradation, harming depending applications, and resulting

Gamze İslamoğlu, Luca Bertaccini, Arpan Suravi Prasad, Francesco Conti

Fast and energy-efficient low-bitwidth floating-point (FP) arithmetic is essential for Artificial Intelligence (AI) systems. Microscaling (MX) standardized formats have recently emerged as a promising alternative to baseline low-bitwidth FP formats, offering improved accuracy with a block-wise shared exponent scale combined with per-element values. However,

Pedro Otero-García, David Pérez-Castro, Manuel Fernández-Veiga, Ana Fernández-Vilas

The advancement of quantum computing threatens classical cryptographic methods, necessitating the development of secure quantum key distribution (QKD) solutions for QKD Networks (QKDN). In this paper, a novel key distribution protocol, Onion Routing Relay (ORR), that integrates onion routing (OR) with post-quantum cryptography (PQC) in a key-relay (KR) model

Yassine El Boudouri, Walter Nuninger, Julian Alvarez, Yvan Peter

Large Language Models (LLMs) demonstrate a notable capacity for adopting personas and engaging in role-playing. However, evaluating this ability presents significant challenges, as human assessments are resource-intensive and automated evaluations can be biased. To address this, we introduce Role-Playing Eval (RPEval), a novel benchmark designed to assess LL

Zhi Liu, Tao Yang, Jing Wang, Yexin Chen

Natural medicines, particularly Traditional Chinese Medicine (TCM), are gaining global recognition for their therapeutic potential in addressing human symptoms and diseases. TCM, with its systematic theories and extensive practical experience, provides abundant resources for healthcare. However, the effective application of TCM requires precise syndrome diag

Liu Jisheng, Zhang Jing

In this paper, we establish the It\^o-Wentzell-Lions formulae for flows of both full and conditional measures on general semimartingales. This generalizes the existing works on flows of measures on It\^o processes. The key technical components involve an appropriate approximation of random fields by cylindrical functions and localization techniques. Moreover

Andreas Holm Akselsen

This paper advances the development of the conformally mapped model for accurate simulation of two-dimensional water waves, here with emphasis on mapping boundaries that represent piston- and flap-type wavemakers. With this, a complete numerical representation of wave flumes is provided -- the first of its kind based on conformal mapping. The model is valida

Philip Groneberg, Saskia Nuñez von Voigt, Thomas Janke, Louis Loechel

In this paper, we present Prink, a novel and practically applicable concept and fully implemented prototype for ks-anonymizing data streams in real-world application architectures. Building upon the pre-existing, yet rudimentary CASTLE scheme, Prink for the first time introduces semantics-aware ks-anonymization of non-numerical (such as categorical or hierar

Jonas Brenig, Radu Timofte

Denoising diffusion models achieved impressive results on several image generation tasks often outperforming GAN based models. Recently, the generative capabilities of diffusion models have been employed for perceptual image compression, such as in CDC. A major drawback of these diffusion-based methods is that, while producing impressive perceptual quality i

Fumihiro Ueno

We classify homogeneous pseudo-Riemannian structures of a three-parameter family of metrics called Kaluza-Klein type on the three-dimensional anti-de Sitter spacetime with their induced groups of isometries and reductive decompositions. We also obtain the classification of homogeneous almost contact and paracontact metric structures of metrics of Kaluza-Klei

Maksim Bobrin, Ilya Zisman, Alexander Nikulin, Vladislav Kurenkov

Behavioral Foundation Models (BFMs) proved successful in producing policies for arbitrary tasks in a zero-shot manner, requiring no test-time training or task-specific fine-tuning. Among the most promising BFMs are the ones that estimate the successor measure learned in an unsupervised way from task-agnostic offline data. However, these methods fail to react

Fang Fang, Christiana Mavroyiakoumou, Leif Ristroph, Michael J. Shelley

We examine theoretically the flow interactions and forward flight dynamics of tandem or in-line flapping wings. Two wings are driven vertically with prescribed heaving-and-plunging motions, and the horizontal propulsion speeds and positions are dynamically selected through aero- or hydro-dynamic interactions. Our simulations employ an improved vortex sheet m

Aswathy Velutharambath, Kai Sassenberg, Roman Klinger

Can deception be detected solely from written text? Cues of deceptive communication are inherently subtle, even more so in text-only communication. Yet, prior studies have reported considerable success in automatic deception detection. We hypothesize that such findings are largely driven by artifacts introduced during data collection and do not generalize be

S. A. Moiseev, K. I. Gerasimov, M. M. Minnegaliev, I. V. Brekotkin

Ramsey interferometry is a cornerstone technique for precise measurement of time and frequency in modern clocks. The Ramsey experiments are typically done in optically dilute samples of atoms to improve homogeneity and avoid back-action of atoms on excitation pulses. In contrast to later belief, we predict and experimentally show that in optically thick samp

Mats Kirkaune, Sven Wedemeyer, Joshiwa van Marrewijk, Tony Mroczkowski

The Atacama Large Millimeter Array (ALMA) has revolutionised the field of solar millimetre astronomy with its high angular resolution and cadence. However, with a limited field of view (FOV), targeted observations of highly dynamic phenomena such of flares are challenging. A large aperture single-dish telescope with a large FOV, such as the future Atacama La

Dongsu Lee, Minhae Kwon

The goal of offline reinforcement learning (RL) is to extract a high-performance policy from the fixed datasets, minimizing performance degradation due to out-of-distribution (OOD) samples. Offline model-based RL (MBRL) is a promising approach that ameliorates OOD issues by enriching state-action transitions with augmentations synthesized via a learned dynam

Haolang Lu, Yilian Liu, Jingxin Xu, Guoshun Nan

The development of Reasoning Large Language Models (RLLMs) has significantly improved multi-step reasoning capabilities, but it has also made hallucination problems more frequent and harder to eliminate. While existing approaches mitigate hallucinations through external knowledge integration, model parameter analysis, or self-verification, they often fail to

Yunhao Ni, Yuxin Guo, Yuhe Liu, Wenxin Sun

This paper studies the approximation capabilities of neural networks that combine layer normalization (LN) with linear layers. We prove that networks consisting of two linear layers with parallel layer normalizations (PLNs) inserted between them (referred to as PLN-Nets) achieve universal approximation, whereas architectures that use only standard LN exhibit

Zheng Wei Lim, Alham Fikri Aji, Trevor Cohn

Large language models (LLMs) are demonstrably capable of cross-lingual transfer, but can produce inconsistent output when prompted with the same queries written in different languages. To understand how language models are able to generalize knowledge from one language to the others, we measure representation similarity between languages, and apply the logit

G. Fotopoulos, N. I. Karachalios, V. Koukouloyannis

Expanding upon our prior findings on the proximity of dynamics between integrable and non-integrable systems within the framework of nonlinear Schr\"odinger equations, we examine this phenomenon for the focusing Discrete Gross-Pitaevskii equation in comparison to the Ablowitz-Ladik lattice. The presence of the harmonic trap necessitates the study of the Ablo

Ilkay Wunderlich, Benjamin Koch, Sven Schönfeld

Convolutional Neural Networks (CNNs) have gained high popularity as a tool for computer vision tasks and for that reason are used in various applications. There are many different concepts, like single shot detectors, that have been published for detecting objects in images or video streams. However, CNNs suffer from disadvantages regarding the deployment on

Emile van Krieken, Pasquale Minervini, Edoardo Ponti, Antonio Vergari

Neurosymbolic (NeSy) predictors combine neural perception with symbolic reasoning to solve tasks like visual reasoning. However, standard NeSy predictors assume conditional independence between the symbols they extract, thus limiting their ability to model interactions and uncertainty - often leading to overconfident predictions and poor out-of-distribution

Robert-Jan Bruintjes, Jan van Gemert

How discriminative position information is for image classification depends on the data. On the one hand, the camera position is arbitrary and objects can appear anywhere in the image, arguing for translation invariance. At the same time, position information is key for exploiting capture/center bias, and scene layout, e.g.: the sky is up. We show that posit

Julia Wunderle, Anton Ehrmanntraut, Jan Pfister, Fotis Jannidis

Encoders remain essential for efficient German NLP and NLU scenarios despite the rise of decoder-only LLMs. This work studies two routes to high-quality German encoders under identical data and training constraints: 1) training from scratch and 2) converting decoders via LLM2Vec. We introduce two resources: ModernGBERT (134M, 1B), fully transparent German en

Bengt Haraldsson, Miroslaw Staron

Context: Large embedded systems in vehicles tend to grow in size and complexity, which causes challenges when maintaining these systems. Objective: We explore how developers perceive the relation between maintainability effort and various sources of complexity. Methods: We conduct a case study at Scania AB, a heavy vehicle OEM. The units of analysis are two

Jop Briët, Davi Castro-Silva

In this paper, we give a quadratic Goldreich-Levin algorithm that is close to optimal in the following ways. Given a bounded function $f$ on the Boolean hypercube $\mathbb{F}_2^n$ and any $\varepsilon>0$, the algorithm returns a quadratic polynomial $q: \mathbb{F}_2^n \to \mathbb{F}_2$ so that the correlation of $f$ with the function $(-1)^q$ is within an ad

Xuejun Guo, Dongxi Ye, Hongbo Yin

Let $E_n$ be the congruent number elliptic curve $y^2=x^3-n^2x$, where $n$ is square-free and not divisible by primes $p\equiv 3\pmod 4$. In this paper, we prove that $L(E_n,1)$ can be expressed as the square of CM values of some simple theta functions, generalizing two classical formulas of Gauss. Our result is meaningful in both theory and practical comput

Antonio Capolupo, Gabriele Pisacane, Aniello Quaranta, Raoul Serao

We analyse the interaction of photons with a scalar dark matter field \phi and we propose to use a single arm interferometer to reveal this interaction and constrain the parameters of the scalar dark matter model. By considering a beam of coherent light and two spatially separated squeezing operations, we show that the interaction of photons with scalar dark

Wei Hua, Chenlin Zhou, Jibin Wu, Yansong Chua

The combination of Spiking Neural Networks (SNNs) with Vision Transformer architectures has garnered significant attention due to their potential for energy-efficient and high-performance computing paradigms. However, a substantial performance gap still exists between SNN-based and ANN-based transformer architectures. While existing methods propose spiking s

Kaede Watanabe, Toshiaki Yachimura, Tsubasa Terada, Hiroshi Kameda

Real-time estimation of ocean wave heights using high-frequency (HF) radar has attracted great attention. This method offers the benefit of easy maintenance by virtue of its ground-based installation. However, it is adversely affected by issues such as low estimation accuracy. As described herein, we propose an algorithm based on the nonnegative sparse regul

Hao Ma, Sabrina Bodmer, Andrea Carron, Melanie Zeilinger

Diffusion models hold great potential in robotics due to their ability to capture complex, high-dimensional data distributions. However, their lack of constraint-awareness limits their deployment in safety-critical applications. We propose Constraint-Aware Diffusion Guidance (CoDiG), a data-efficient and general-purpose framework that integrates barrier func

Muhammad Awais Amin, Adama Ilboudo, Abdul Samad bin Shahid, Amjad Ali

One of the major challenges in the field of computer vision especially for detection, segmentation, recognition, monitoring, and automated solutions, is the quality of images. Image degradation, often caused by factors such as rain, fog, lighting, etc., has a negative impact on automated decision-making.Furthermore, several image restoration solutions exist,

Kevin Chenhao Li, Vahid Zolfaghari, Nenad Petrovic, Fengjunjie Pan

The Object Constraint Language (OCL) is essential for defining precise constraints within Model-Based Systems Engineering (MBSE). However, manually writing OCL rules is complex and time-consuming. This study explores the optimization of Retrieval-Augmented Generation (RAG) for automating OCL rule generation, focusing on the impact of different retrieval stra

Tsz Pang Yuen, Eni Musta, Ingrid Van Keilegom

In survival analysis, estimating the fraction of 'immune' or 'cured' subjects who will never experience the event of interest, requires a sufficiently long follow-up period. A few statistical tests have been proposed to test the assumption of sufficient follow-up, i.e. whether the right extreme of the censoring distribution exceeds that of the survival time

Chiara Fend, Claudia Redenbach

Spatial point processes are used as models in many different fields ranging from ecology and forestry to cosmology and materials science. In recent years, model validation, and in particular goodness-of-fit testing of a proposed point process model have seen many advances. Most of the proposed tests are based on a functional summary statistic of the observed

Liancheng Gong, Wang Zhu, Jesse Thomason, Li Zhang

Using LLMs not to predict plans but to formalize an environment into the Planning Domain Definition Language (PDDL) has been shown to improve performance and control. While most existing methodology only applies to fully observable environments, we adapt to the more realistic and challenging partially observable environments without sufficient information to

A. V. Korybut

In the recently proposed generating systems for the (anti)holomorphic sector of the 4d higher spin theory and for the off-shell higher spin theory in generic dimension locality was achieved due to a peculiar limiting star product. Even though the generating systems exhibit all-order locality, the product itself encounters uncertainties when functions from sp

Francesco Innocenti, El Mehdi Achour, Christopher L. Buckley

The biological implausibility of backpropagation (BP) has motivated many alternative, brain-inspired algorithms that attempt to rely only on local information, such as predictive coding (PC) and equilibrium propagation. However, these algorithms have notoriously struggled to train very deep networks, preventing them from competing with BP in large-scale sett

Di Zhang, Ligang Liu

We present an asymptotic analysis of shell lattice metamaterials based on Ciarlet's shell theory, introducing a new metric--asymptotic directional stiffness (ADS)--to quantify how the geometry of the middle surface governs the effective stiffness. We prove a convergence theorem that rigorously characterizes ADS and establishes its upper bound, along with nec

Jian Liu, Haohan Weng, Biwen Lei, Xianghui Yang

The next-coordinate prediction paradigm has emerged as the de facto standard in current auto-regressive mesh generation methods. Despite their effectiveness, there is no efficient measurement for the various tokenizers that serialize meshes into sequences. In this paper, we introduce a new metric Per-Token-Mesh-Entropy (PTME) to evaluate the existing mesh to

Snehashis Majhi, Giacomo D'Amicantonio, Antitza Dantcheva, Quan Kong

Weakly-supervised methods for video anomaly detection (VAD) are conventionally based merely on RGB spatio-temporal features, which continues to limit their reliability in real-world scenarios. This is due to the fact that RGB-features are not sufficiently distinctive in setting apart categories such as shoplifting from visually similar events. Therefore, tow

François Bachoc, Jérôme Bolte, Ryan Boustany, Jean-Michel Loubes

Despite growing empirical evidence of bias amplification in machine learning, its theoretical foundations remain poorly understood. We develop a formal framework for majority-minority learning tasks, showing how standard training can favor majority groups and produce stereotypical predictors that neglect minority-specific features. Assuming population and va

E. Contreras, A. Di Teodoro, M. Mena

In this work, we introduce a new fractional derivative that modifies the conventional Riemann-Liouville operator to obtain a set of fractional Einstein field equations within a 2+1 dimensional spacetime by assuming a static and circularly symmetric metric. The main reason for introducing this new derivative stems from addressing the divergence encountered du

Paul Pöhl, Viktor Schlegel, Hao Li, Anil Bharath

Generating synthetic ECG data has numerous applications in healthcare, from educational purposes to simulating scenarios and forecasting trends. While recent diffusion models excel at generating short ECG segments, they struggle with longer sequences needed for many clinical applications. This paper proposes a novel three-layer synthesis framework for genera

Alessio Cela, Aleksander Doan

We prove a symplectic version of a conjecture of Lian and Pandharipande: in sufficiently high degree, the fixed-domain Gromov-Witten invariants of positive symplectic manifolds are signed counts of pseudo-holomorphic curves. The original conjecture in the complex algebraic setting was recently disproved by Beheshti et al. However, we show that the statement

Andrew Mummery, Jiachen Jiang, Adam Ingram, Andrew Fabian

Emission from within the plunging region of black hole accretion flows has recently been detected in two X-ray binary systems. There is, furthermore, a possible discrepancy between the inferred spins of gravitational wave and electromagnetically detected black holes. Motivated by these two results we demonstrate, using theoretical calculations, numerical sim

Marouane Il Idrissi, Agathe Fernandes Machado, Ewen Gallic, Arthur Charpentier

Cooperative game theory methods, notably Shapley values, have significantly enhanced machine learning (ML) interpretability. However, existing explainable AI (XAI) frameworks mainly attribute average model predictions, overlooking predictive uncertainty. This work addresses that gap by proposing a novel, model-agnostic uncertainty attribution (UA) method gro

Tal Agranov, Robert L. Jack, Michael E. Cates, Étienne Fodor

We study the entropy production rate (EPR) of aligning self-propelled particles which undergo a flocking transition towards a polarized collective motion. In our thermodynamically consistent lattice model, individual self-propulsion is the exclusive source of irreversibility. We derive the fluctuating hydrodynamics for large system sizes using a controlled c

Kathrin Lammers, Valerie Vaquet, Barbara Hammer

As machine learning is increasingly applied in an online fashion to deal with evolving data streams, the fairness of these algorithms is a matter of growing ethical and legal concern. In many use cases, class imbalance in the data also needs to be dealt with to ensure predictive performance. Current fairness-aware stream learners typically attempt to solve t

Debarpan Bhattacharya, Apoorva Kulkarni, Sriram Ganapathy

The popular success of text-based large language models (LLM) has streamlined the attention of the multimodal community to combine other modalities like vision and audio along with text to achieve similar multimodal capabilities. In this quest, large audio language models (LALMs) have to be evaluated on reasoning related tasks which are different from tradit

Prosper Rosaire Mama Assandje, Joseph Dongho, Thomas Bouetou Bouetou

In this paper, we describe the evolution of spectral curves in the Siegel Jacobi space through the Schrodinger equation constructed from a Kahler geometry induced on the lognormal statistical manifold via Dombrowski's construction. We introduce new holomorphic structures and show that the Hamiltonian vector field coincides with the fundamental vector field g

Karel Proesmans, Gianmaria Falasco, Atul Tanaji Mohite, Massimiliano Esposito

Aligning self-propelled particles undergo a nonequilibrium flocking transition from apolar to polar phases as their interactions become stronger. We propose a thermodynamically consistent lattice model, in which the internal state of the particles biases their diffusion, to capture such a transition. Changes of internal states and jumps between lattice sites

Zi-Ying Chen, Chuan-Xian Ren, Hong Yan

Partial domain adaptation (PDA) problem requires aligning cross-domain samples while distinguishing the outlier classes for accurate knowledge transfer. The widely used weighting framework tries to address the outlier classes by introducing the reweighed source domain with a similar label distribution to the target domain. However, the empirical modeling of

Rodrigo Maulen-Soto, Pierre Marion, Claire Boyer

Transformers have emerged as a powerful neural network architecture capable of tackling a wide range of learning tasks. In this work, we provide a theoretical analysis of their ability to automatically extract structure from data in an unsupervised setting. In particular, we demonstrate their suitability for clustering when the input data is generated from a

Sungmin Cha, Kyunghyun Cho

Knowledge distillation (KD) is a core component in the training and deployment of modern generative models, particularly large language models (LLMs). While its empirical benefits are well documented -- enabling smaller student models to emulate the performance of much larger teachers -- the underlying mechanisms by which KD improves generative quality remai

Zexin Hu, Yong Gao, Lijing Shao

Neutron stars (NSs) are excellent laboratories for testing gravity theories as they are strongly self-gravitating bodies and have rich observational phenomena. However, strong-field gravity effects in NS could be degenerate with their equation of state (EOS) which is largely unknown. Fortunately, there exist the so-called universal relations among the NS mac

Guangda Liu, Chengwei Li, Zhenyu Ning, Jing Lin

Large language models (LLMs) are widely deployed with rapidly expanding context windows to support increasingly demanding applications. However, long contexts pose significant deployment challenges, primarily due to the KV cache whose size grows proportionally with context length. While KV cache compression methods have been proposed to address this issue, K

Saurabh Shrivastava, Kalachand Shuin

For $f,g \in \mathscr{S}(\R^n), n\geq 3$, consider the bilinear cone multiplier operator defined by \[{T}^{\lambda}_{R}(f,g)(x):=\int_{\mathbb{R}^{2n}}m^{\lambda}\left(\frac{\xi'}{R\xi_n},\frac{\eta'}{R\eta_n}\right)\hat{f}(\xi)\hat{g}(\eta)e^{2\pi\iota x\cdot(\xi+\eta)}~d\xi d\eta,\] where $\lambda>0, R>0$ and \[m^{\lambda}\left(\frac{\xi'}{R\xi_n},\frac{\e

Gabriel Elyas Gama Araujo, Andreia Luisa da Rosa, Alexandre Cavalheiro Dias, Thomas Frauenheim

In this work we use first principles density-functional theory and Bethe-Salpeter equation together with tight-binding based maximally localized wannier functions (MLWF-TB) to investigate the electronic, optical and topological properties of two-dimensional bismuth (bismuthene) containing vacancy defects. We demonstrate that these properties depends on the s

László Csató

The organisers of major sports competitions use different policies with respect to constraints in the group draw. Our paper aims to rationalise these choices by analysing the trade-off between attractiveness (the number of games played by teams from the same geographic zone) and fairness (the departure of the draw mechanism from a uniform distribution). A pa

Ram Padmanabhan, Antoine Aspeel, Necmiye Ozay, Melkior Ornik

In this paper, we consider the problem of designing prefix-based optimal controllers for switched linear systems over finite horizons. This problem arises in fault-tolerant control, when system faults result in abrupt changes in dynamics. We consider a class of mode-prefix-based linear controllers that depend only on the history of the switching signal. The

Ahmed Boughdiri, Clément Berenfeld, Julie Josse, Erwan Scornet

Generalization methods offer a powerful solution to one of the key drawbacks of randomized controlled trials (RCTs): their limited representativeness. By enabling the transport of treatment effect estimates to target populations subject to distributional shifts, these methods are increasingly recognized as the future of meta-analysis, the current gold standa

Han Zheng, Ilia Shumailov, Tianqi Fan, Aiden Hall

The rapid advancement of bug-finding techniques has led to the discovery of more vulnerabilities than developers can reasonably fix, creating an urgent need for effective Automated Program Repair (APR) methods. However, the complexity of modern bugs often makes precise root cause analysis difficult and unreliable. To address this challenge, we propose crash-

Shaowu Wu, Liting Zeng, Wei Lu, Xiangyang Luo

With the rapid rise of large models, copyright protection for generated image content has become a critical security challenge. Although deep learning watermarking techniques offer an effective solution for digital image copyright protection, they still face limitations in terms of visual quality, robustness and generalization. To address these issues, this

Christodoulos Kechris, Jonathan Dan, David Atienza

Traditional saliency map methods, popularized in computer vision, highlight individual points (pixels) of the input that contribute the most to the model's output. However, in time series, they offer limited insights, as semantically meaningful features are often found in other domains. We introduce Cross-domain Integrated Gradients, a generalization of Inte