October 2024 arXiv papers — page 8

Andrea Beraudo

Recent results on particle production in hadronic collisions at the LHC, from proton-proton (pp) to nucleus-nucleus (AA), challenge the traditional paradigm of hadronization as a universal late-time process which can be factorized from the partonic description of the rest of the event. If the violation of this description in nuclear collisions has been accep

Dizhen Liang

Transformer-based architectures have achieved remarkable success in natural language processing and computer vision. However, their performance in multivariate long-term forecasting often falls short compared to simpler linear baselines. Previous research has identified the traditional attention mechanism as a key factor limiting their effectiveness in this

Xuyuan Liu, Yinghao Cai, Qihui Yang, Yujun Yan

Graph Neural Networks (GNNs) have emerged as a dominant approach in graph representation learning, yet they often struggle to capture consistent similarity relationships among graphs. While graph kernel methods such as the Weisfeiler-Lehman subtree (WL-subtree) and Weisfeiler-Lehman optimal assignment (WLOA) kernels are effective in capturing similarity rela

Rajmeet Singh, lakmal Seneviratne, Irfan Hussain

The decline of bee and wind-based pollination systems in greenhouses due to controlled environments and limited access has boost the importance of finding alternative pollination methods. Robotic based pollination systems have emerged as a promising solution, ensuring adequate crop yield even in challenging pollination scenarios. This paper presents a compre

Junchao Wu, Runzhe Zhan, Derek F. Wong, Shu Yang

Detecting text generated by large language models (LLMs) is of great recent interest. With zero-shot methods like DetectGPT, detection capabilities have reached impressive levels. However, the reliability of existing detectors in real-world applications remains underexplored. In this study, we present a new benchmark, DetectRL, highlighting that even state-o

Yongqi Zhuo, Zhengyuan Su, Chenggang Zhao, Mingyu Gao

The desires for better prediction accuracy and higher execution performance in neural networks never end. Neural architecture search (NAS) and tensor compilers are two popular techniques to optimize these two goals, but they are both limited to composing or optimizing existing manually designed operators rather than coming up with completely new designs. In

Sarina Thomas, Qing Cao, Anna Novikova, Daria Kulikova

Ejection fraction (EF) of the left ventricle (LV) is considered as one of the most important measurements for diagnosing acute heart failure and can be estimated during cardiac ultrasound acquisition. While recent successes in deep learning research successfully estimate EF values, the proposed models often lack an explanation for the prediction. However, pr

Ming Li, Yanhong Li, Tianyi Zhou

What makes a difference in the post-training of LLMs? We investigate the training patterns of different layers in large language models (LLMs) through the lens of the gradient. We are specifically interested in how fast vs. slow thinking affects the layer-wise gradients, given the recent popularity of training LLMs on reasoning paths such as chain-of-thought

Karim Kassab, Antoine Schnepf, Jean-Yves Franceschi, Laurent Caraffa

Tri-Planar NeRFs enable the application of powerful 2D vision models for 3D tasks, by representing 3D objects using 2D planar structures. This has made them the prevailing choice to model large collections of 3D objects. However, training Tri-Planes to model such large collections is computationally intensive and remains largely inefficient. This is because

John Ben S Temones

This study explores the link between dietary behavior and the risk of nutritional deficiency dermatoses (NDD) in the Bicol region, where malnutrition remains a concern. Using regression analysis on FNRI data, it examines food purchase patterns, particularly riboflavin intake. Findings show an NDD risk prevalence of 15.75%, with Masbate and Camarines Sur cont

E. S. Carrera, Y. Zhang, J-D. Bancal, N. Sangouard

The Wineland parameter aims at detecting metrologically useful entangled states, called spin-squeezed states, from expectations and variances of total angular momenta. {However, efficient strategies for estimating this parameter in practice have yet to be determined and in particular, the effects of a finite number of measurements remain insufficiently addre

Sami Räsänen

Large experimental Collaborations at the LHC (ALICE, ATLAS, CMS, and LHCb) bring together over 13,000 people from hundreds of institutes over the world. There are many (working) cultures inside these international Collaborations. It is important to acknowledge that cultural differences exist and manifest in our workspaces in various ways such as communicatio

Niclas Krupp, Gerrit Groenhof, Oriol Vendrell

Strong coupling between excitons and confined modes of light presents a promising pathway to tunable and enhanced energy transport in organic materials. By forming hybrid light-matter quasiparticles, exciton-polaritons, electronic excitations can traverse long distances at high velocities through ballistic flow. However, transport behavior of exciton-polarit

Yufeng Jiang, Zongxi Li, Xiangyan Chen, Haoran Xie

Recent advancements in medical imaging have resulted in more complex and diverse images, with challenges such as high anatomical variability, blurred tissue boundaries, low organ contrast, and noise. Traditional segmentation methods struggle to address these challenges, making deep learning approaches, particularly U-shaped architectures, increasingly promin

JaeYoon Kim, Junyu Xuan, Christy Liang, Farookh Hussain

Offline-to-online reinforcement learning (RL) leverages both pre-trained offline policies and online policies trained for downstream tasks, aiming to improve data efficiency and accelerate performance enhancement. An existing approach, Policy Expansion (PEX), utilizes a policy set composed of both policies without modifying the offline policy for exploration

Haiwen Li, Fei Su, Zhicheng Zhao

As a challenging vision-language task, Zero-Shot Composed Image Retrieval (ZS-CIR) is designed to retrieve target images using bi-modal (image+text) queries. Typical ZS-CIR methods employ an inversion network to generate pseudo-word tokens that effectively represent the input semantics. However, the inversion-based methods suffer from two inherent issues: Fi

Debashish Chowdhury, Andreas Schadschneider, Katsuhiro Nishinari

Enormous progress have been made in the last 20 years since the publication of our review \cite{csk05polrev} in this journal on transport and traffic phenomena in biology. In this brief article we present a glimpse of the major advances during this period. First, we present similarities and differences between collective intracellular transport of a single m

Cihan Okay, Atak Talay Yucel, Selman Ipek

We introduce a new classical simulation algorithm based on non-signaling polytopes of multipartite Bell scenarios, capable of simulating universal measurement-based quantum computation with single-qubit Pauli measurements. In our model, the simultaneous presence of non-stabilizerness and entanglement is necessary for quantum speedup. The region of quantum st

Silvia Cingolani, Marco Gallo, Norihisa Ikoma, Kazunaga Tanaka

We study the following nonlinear Schr\"odinger equation and we look for normalized solutions $(\mu,u)\in {\bf R}\times H^1({\bf R}^N)$ for a given $m>0$ and $N\geq 2$ \[ -\Delta u + \mu u = g(u)\quad \text{in}\ {\bf R}^N, \qquad \frac{1}{2}\int_{{\bf R}^N} u^2 dx = m. \] We assume that $g$ has an $L^2$-critical growth, both at the origin and at infinity. Tha

Guanhua Liu

In this paper we give equivalent conditions for the weak parabolic Harnack inequality for general regular Dirichlet forms without killing part, in terms of local heat kernel estimates or growth lemmas. With a tail estimate on the jump measure, we obtain from these conditions the H\"older continuity of caloric and harmonic functions. Our results generalize th

Gérard Henry Edmond Duchamp, Joseph Ben Geloun, Christophe Tollu

We show that the algebra of the coloured rook monoid $R_n^{(r)}$, {\em i.e.} the monoid of $n \times n$ matrices with at most one non-zero entry (an $r$-th root of unity) in each column and row, is the algebra of a finite groupoid, thus is endowed with a $C^*$-algebra structure. This new perspective uncovers the representation theory of these monoid algebras

Yuki Hirakawa, Takashi Wada, Kazuya Morishita, Ryotaro Shimizu

Fashion aesthetic evaluation is the task of estimating how well the outfits worn by individuals in images suit them. In this work, we examine the zero-shot performance of GPT-4V on this task for the first time. We show that its predictions align fairly well with human judgments on our datasets, and also find that it struggles with ranking outfits in similar

Pierre-Alexandre Bliman

Sterile Insect Technique (SIT) is a promising control method against insect pests and insect vectors. It consists in releasing males previously sterilized in laboratory, in order to reduce or eliminate a specific wild population. We study in this paper the implementation by feedback control of SIT-based elimination campaign of Aedes mosquitoes. We provide st

Georges Gagneré

We propose to review the main stages in the computer history of virtual actors, with a view to the exploration of virtual reality and discussion on different approaches to human simulation. The notion of autonomy emerges as a key issue for the virtual entities. We then explore one way of building elements of autonomy and conclude with an example of avatar st

Irina Tolstykh, Aleksandra Tsybina, Sergey Yakubson, Aleksandr Gordeev

With the increasing quality and spread of LLM assistants, the amount of generated content is growing rapidly. In many cases and tasks, such texts are already indistinguishable from those written by humans, and the quality of generation continues to increase. At the same time, detection methods are advancing more slowly than generation models, making it chall

Yaowei Xie, Huan Yu

In this paper, we establish the mild ill-posedness of 2D IPM equation in the critical Sobolev space $W^{1,\infty}$ when the initial data are small perturbations of stable profile $g(x_2).$ Consequently, instability can be inferred. Notably, our results are valid for arbitrary vertically stratified density profiles $g(x_2)$ without imposing any restrictions o

Debangshu Banerjee, Aditya Gopalan

Recent advances in aligning Large Language Models with human preferences have benefited from larger reward models and better preference data. However, most of these methodologies rely on the accuracy of the reward model. The reward models used in Reinforcement Learning with Human Feedback (RLHF) are typically learned from small datasets using stochastic opti

Taridzo Chomutare, Therese Olsen Svenning, Miguel Ángel Tejedor Hernández, Phuong Dinh Ngo

\textbf{Trial design} Crossover randomized controlled trial. \textbf{Methods} An AI tool, Easy-ICD, was developed to assist clinical coders and was tested for improving both accuracy and time in a user study in Norway and Sweden. Participants were randomly assigned to two groups, and crossed over between coding complex (longer) texts versus simple (shorter)

Antonio Rago, Kristijonas Čyras, Jack Mumford, Oana Cocarascu

This chapter provides an overview of research works that present approaches with some degree of cross-fertilisation between Computational Argumentation and Machine Learning. Our review of the literature identified two broad themes representing the purpose of the interaction between these two areas: argumentation for machine learning and machine learning for

Yan zilong, Peng Yuejian

For a family $\mathcal{F}$ of subsets of a finite set, define $\mathcal{D}(\mathcal{F})=\{F\setminus F': F, F'\in\mathcal{F}\}$. A family $\mathcal{F}$ is called intersecting if $F\cap F'\not=\emptyset$ for all $F, F'\in\mathcal{F}$. Frankl \cite{Frankl} showed that for a $k$-uniform intersecting family $\mathcal{F}\subset{[n]\choose k}$ with $n\ge k(k+3)$,

Andrea Giordano, Luca Romanello, Diego Perez Gonzalez, Mirko Kovac

This paper underscores the importance of environmental monitoring, and specifically of freshwater ecosystems, which play a critical role in sustaining life and global economy. Despite their importance, insufficient data availability prevents a comprehensive understanding of these ecosystems, thereby impeding informed decision-making concerning their preserva

Ziyu Ye, Rishabh Agarwal, Tianqi Liu, Rishabh Joshi

Current reinforcement learning (RL) frameworks for large language models (LLM) post-training typically assume a fixed prompt distribution, which is sub-optimal and bottlenecks scalability. Prior works have explored prompt evolving, but are often limited to the supervised fine-tuning stage, and prompts are sampled and evolved uniformly without signals. This e

Oliver Hahn, Maxime Garnier, Giulia Ferrini, Alessandro Ferraro

State conversion is a fundamental task in quantum information processing. Quantum resource theories allow for analyzing and bounding conversions that use restricted sets of operations. In the context of continuous-variable systems, state conversions restricted to Gaussian operations are crucial for both fundamental and practical reasons, particularly in stat

Georg C. Hofstätter, Philipp Kniefacz, Franz E. Schuster

It is shown that each continuous even Minkowski valuation on convex bodies of degree $1 \leq i \leq n - 1$ intertwining rigid motions is obtained from convolution of the $i$th projection function with a unique spherical Crofton distribution. In case of a non-negative distribution, the polar volume of the associated Minkowski valuation gives rise to an isoper

Rui-Cheng Guo, Yanwu Gu, Dong E. Liu

Simulating quantum many-body systems is crucial for advancing physics but poses substantial challenges for classical computers. Quantum simulations overcome these limitations, with analog simulators offering unique advantages over digital methods, such as lower systematic errors and reduced circuit depth, making them efficient for studying complex quantum ph

Xiufeng Huang, Ruiqi Li, Yiu-ming Cheung, Ka Chun Cheung

3D Gaussian Splatting (3DGS) has become a crucial method for acquiring 3D assets. To protect the copyright of these assets, digital watermarking techniques can be applied to embed ownership information discreetly within 3DGS models. However, existing watermarking methods for meshes, point clouds, and implicit radiance fields cannot be directly applied to 3DG

Isshin Sumiyoshi, Yoshitaro Nose

The steady-state distribution of carrier concentrations in a solar cell under operating conditions is a key source of carrier recombination and directly influences the output current density. In this study, we investigated the effects of illumination and bias voltage on the steady-state distributions of carrier concentrations and recombination rates in a hom

Silvia Musolino, Gianni Aupetit-Diallo, Mathias Albert, Patrizia Vignolo

We investigate the dynamics of one-dimensional SU(2) ultracold fermions near the Tonks-Girardeau limit, confined in a box potential. The system is driven out of equilibrium by initially preparing the two spin components in a fully separated configuration, and its evolution is described by the Hamiltonian in the presence of strong repulsive interactions. Buil

Jia Song, Wanru Zhuang, Yujie Lin, Liang Zhang

Cross-modal text-molecule retrieval model aims to learn a shared feature space of the text and molecule modalities for accurate similarity calculation, which facilitates the rapid screening of molecules with specific properties and activities in drug design. However, previous works have two main defects. First, they are inadequate in capturing modality-share

Prabhat Kumar, Roger A Sauer, Anupam Saxena

This paper presents a topology optimization approach to design 2D contact-aided compliant mechanisms (CCMs) that can trace the desired output paths with more than one kink while experiencing self and/or external contacts. Such CCMs can be used as mechanical compliant switches. Hexagonal elements are used to parameterize the design domain. Negative circular m

Georgi Vodev

In this paper we extend the results in [16] to more general domains. More precisely, we obtain transmission eigenvalue-free regions for the interior transmission problem with one complex-valued refraction index, that is, with a damping term which does not vanish on the boundary. In particular, we remove the non-trapping condition as well as the strict concav

Ugo de Noyers, Maud Sarazin, Björn Herrmann

We present an extensive phenomenological study of a scotogenic framework including a scalar singlet, a scalar doublet, a fermionic doublet, and two generations of a fermionic triplet, allowing to provide three non-zero neutrino masses and three viable dark matter candidates. Using a Markov Chain Monte Carlo numerical technique, we probe the parameter space o

Alex Adam, Carl Blair, Chunnong Zhao

The next generation of gravitational wave detectors will move to cryogenic operation in order to reduce thermal noise and thermal distortion. This necessitates a change in mirror substrate with silicon being a good candidate. Birefringence is an effect that will degrade the sensitivity of a detector and is of greater concern in silicon due to its crystalline

Vishnu Muraleedharan Sajitha, Bodhaditya Santra, Matthew J. Davis, L. A. Williamson

We identify and interpret the possible quantum thermal machine regimes with a transverse-field Ising model as the working substance. In general, understanding the emergence of such regimes in a many-body quantum system is challenging due to the dependence on the many energy levels in the system. By considering infinitesimal work strokes, we can understand th

Carlos Duaso Pueyo, Harry Goodhew, Ciaran McCulloch, Enrico Pajer

Physical theories have a limited regime of validity and hence must be accompanied by a breakdown diagnostic to establish when they cease to be valid as parameters are varied. For perturbative theories, estimates of the first neglected order offer valuable guidance, but one is often interested in sharp bounds beyond which perturbation theory necessarily fails

Nouh Zeggai, Martino LoBue, Morgan Almanza

In response to the growing demand for more efficient and compact refrigeration and energy conversion devices, electrocaloric poly(vinylidene fluoride-trifluoroethylene-chlorofluoroethylene) is among the most promising active substances. However, despite its high electrocaloric response, the maximum efficiency achievable over a cooling cycle is hampered by lo

Kemal Firdaus, Jörn Behrens

We formulate a depth-averaged non-hydrostatic model to solve wave equations with generation by a moving bottom. This model is built upon the shallow water equations, which are widely used in tsunami wave modelling. An extension leads to two additional unknowns to be solved: vertical momentum and non-hydrostatic pressure. We show that a linear vertical veloci

Lujia Bai, David Veitch, Weichi Wu, Wenyang Zhang

This paper studies high-dimensional trend inference for piecewise smooth signals under nonstationary noise and asynchronous structural breaks by first detecting asynchronous changes without assuming stationarity and then further exploiting latent group structures to estimate trend functions. In the first step, we propose AJDN (Asynchronous Jump Detection und

Laura Barros, Aiko Schmeißer

Adverse economic shocks are known to reshape voter behavior -- the demand side of politics. Much less is known about their consequences for the supply side: how such shocks affect who becomes a politician. This paper examines how job losses influence individuals' decisions to enter politics and the implications for political selection. Using administrative d

Kimon Kieslich, Nicholas Diakopoulos, Natali Helberger

Impact assessments have emerged as a common way to identify the negative and positive implications of AI deployment, with the goal of avoiding the downsides of its use. It is undeniable that impact assessments are important - especially in the case of rapidly proliferating technologies such as generative AI. But it is also essential to critically interrogate

Junda Wu, Xintong Li, Ruoyu Wang, Yu Xia

Offline evaluation of LLMs is crucial in understanding their capacities, though current methods remain underexplored in existing research. In this work, we focus on the offline evaluation of the chain-of-thought capabilities and show how to optimize LLMs based on the proposed evaluation method. To enable offline feedback with rich knowledge and reasoning pat

Chengli Li, Yurui Tang, Xingzhi Zhan

A local subgraph of a graph is the subgraph induced by the neighborhood of a vertex. Thus a graph of order $n$ has $n$ local subgraphs. A graph $G$ is called locally nonforesty if every local subgraph of $G$ contains a cycle. Recently, in studying forest cuts of a graph, Chernyshev, Rauch and Rautenbach posed the conjecture that if $n$ and $m$ are the order

Tyler Ga Wei Lum, Albert H. Li, Preston Culbertson, Krishnan Srinivasan

This work explores conditions under which multi-finger grasping algorithms can attain robust sim-to-real transfer. While numerous large datasets facilitate learning generative models for multi-finger grasping at scale, reliable real-world dexterous grasping remains challenging, with most methods degrading when deployed on hardware. An alternate strategy is t

Vincent Andrieu, Daniele Astolfi, Alexandre Cellier-Devaux

In this note we investigate the problem of global exponential synchronization of multi-agent systems described by nonlinear input affine dynamics. We consider the case of networks described by undirected connected graphs possibly without leader. We present a set of sufficient conditions based on a Riemannian metric approach in order to design a state-feedbac

Zhu-yao Jin, Jun Jing

In this paper, we derive universal nonadiabatic passages in a general $M+N$-dimensional discrete system, where $M$ and $N$ denote the degrees of freedom for the assistant and working subspaces, respectively, that could be separated by rotation or energy and coupled through driving. A systematic method is provided to construct parametric ancillary bases by th

Chen Huang, Skyler Seto, Samira Abnar, David Grangier

Large pretrained vision-language models like CLIP have shown promising generalization capability, but may struggle in specialized domains (e.g., satellite imagery) or fine-grained classification (e.g., car models) where the visual concepts are unseen or under-represented during pretraining. Prompt learning offers a parameter-efficient finetuning framework th

V. G. Bordo

The first-principles theory of lasing in a rectangular lattice of spherical metal nanoparticles is developed in a fully analytical form in the dipole approximation. The lasing conditions are obtained for different diffraction orders, both propagating and evanescent. Their analysis reveals that besides usual lasing there can be lasing without population inver

E. M. Glaviano, A. Bonanno

We study the flow of the non-local truncation in quantum gravity and we focus in particular on the Polyakov effective action for a non-minimally coupled scalar field on a two dimensional curved space. We show that it is possible to explicitly integrate the flow of all the local and non-local operator terms up to $k=0$ and recover effective action without the

Chenxin Tu, Xiaowei Cui, Gang Liu, Sihao Zhao

In a time-division broadcast positioning system (TDBPS), localizing mobile targets using classical time difference of arrival (TDOA) methods poses significant challenges. Concurrent TDOA measurements are infeasible because targets receive signals from different anchors and extract their transmission times at different reception times, as well as at varying p

O. A. Dobush, M. P. Kozlovskii, R. V. Romanik

A cell fluid model with a modified Morse potential is studied. The supercritical states are considered with respect to a possibility to build a separation boundary between liquid-like and gas-like bahaviors. Three different lines are calculated that can be used for this purpose: the locus of the isothermal compressibility maxima, the locus of the thermal exp

Wenhan Chang, Tianqing Zhu, Ping Xiong, Yufeng Wu

In the rapid advancement of artificial intelligence, privacy protection has become crucial, giving rise to machine unlearning. Machine unlearning is a technique that removes specific data influences from trained models without the need for extensive retraining. However, it faces several key challenges, including accurately implementing unlearning, ensuring p

Peizhi Tang, Chuang Yang, Tong Xing, Xiaohang Xu

Human mobility prediction plays a critical role in applications such as disaster response, urban planning, and epidemic forecasting. Traditional methods often rely on designing crafted, domain-specific models, and typically focus on short-term predictions, which struggle to generalize across diverse urban environments. In this study, we introduce Llama3-8B-M

Samuel Holt, Tennison Liu, Mihaela van der Schaar

Digital Twins (DTs) are computational models that simulate the states and temporal dynamics of real-world systems, playing a crucial role in prediction, understanding, and decision-making across diverse domains. However, existing approaches to DTs often struggle to generalize to unseen conditions in data-scarce settings, a crucial requirement for such models

Mahin Mohammadi, Saman Jamshidi

Brain tumors pose a serious health threat due to their rapid growth and potential for metastasis. While medical imaging has advanced significantly, accurately identifying and characterizing these tumors remains a challenge. This study addresses this challenge by leveraging the innovative TrAdaBoost methodology to enhance the Brain Tumor Segmentation (BraTS20

Xiaomeng Wang, Nan Wang, Guofeng Zhang

In this paper, we propose a flexible SLAM framework, XRDSLAM. It adopts a modular code design and a multi-process running mechanism, providing highly reusable foundational modules such as unified dataset management, 3d visualization, algorithm configuration, and metrics evaluation. It can help developers quickly build a complete SLAM system, flexibly combine

Jin-Woo Kim, Suseong Lim, Heonoh Kim, June Koo Kevin Rhee

Verifying the violation of Bell's inequality is one of the most representative methods to demonstrate that entangled photon pairs prepared in a quantum optics-based system exhibit quantum properties. While experiments on Bell inequality violations have been theoretically well-established and extensively conducted to implement various quantum information tech

Aran Chakraborty, Yushi Tsukimoto, August Posch, Jack Watson

The resilience of urban rapid transit systems (URTs) to a rapidly evolving threat space is of much concern. Extreme rainfall events are both intensifying and growing more frequent under continuing climate change, exposing transit systems to flooding, while cyber threats and emerging technologies such as unmanned aerial vehicles are exposing such systems to t

Chiyu Zhang, Lu Zhou, Xiaogang Xu, Jiafei Wu

With the advent of Large Vision-Language Models (LVLMs), new attack vectors, such as cognitive bias, prompt injection, and jailbreaking, have emerged. Understanding these attacks promotes system robustness improvement and neural networks demystification. However, existing surveys often target attack taxonomy and lack in-depth analysis like 1) unified insight

Junshu Sun, Chenxue Yang, Xiangyang Ji, Qingming Huang

Message passing plays a vital role in graph neural networks (GNNs) for effective feature learning. However, the over-reliance on input topology diminishes the efficacy of message passing and restricts the ability of GNNs. Despite efforts to mitigate the reliance, existing study encounters message-passing bottlenecks or high computational expense problems, wh

Usman Hasan, A. Uzair, M. Junaid, Bilal Tariq

The rare three body decay of a Higgs boson to a lepton-anti lepton pair and a photon has begun to attract attention, after the first evidence for the $H\to Z\gamma$ at CMS and ATLAS, which is a sub process of $H \to \ell^+ \ell^- \gamma$ . To investigate some important features of this process, we suggest that the polarized forward-backward and the photon po

Eugene Jang, Kimin Lee, Jin-Woo Chung, Keuntae Park

Tokenization is a crucial step that bridges human-readable text with model-readable discrete tokens. However, recent studies have revealed that tokenizers can be exploited to elicit unwanted model behaviors. In this work, we investigate incomplete tokens, i.e., undecodable tokens with stray bytes resulting from byte-level byte-pair encoding (BPE) tokenizatio

Fan Yao, Yiming Liao, Jingzhou Liu, Shaoliang Nie

On User-Generated Content (UGC) platforms, recommendation algorithms significantly impact creators' motivation to produce content as they compete for algorithmically allocated user traffic. This phenomenon subtly shapes the volume and diversity of the content pool, which is crucial for the platform's sustainability. In this work, we demonstrate, both theoret

Shintaro Inoue, Kento Kawaharazuka, Temma Suzuki, Sota Yuzaki

Humanoids exhibit a wide variety in terms of joint configuration, actuators, and degrees of freedom, resulting in different achievable movements and tasks for each type. Particularly, musculoskeletal humanoids are developed to closely emulate human body structure and movement functions, consisting of a skeletal framework driven by numerous muscle actuators.

Juan Zhang, Junyue Luo

In this paper, we apply the practical GADI-HS iteration as a smoother in algebraic multigrid (AMG) method for solving second-order non-selfadjoint elliptic problem. Additionally, we prove the convergence of the derived algorithm and introduce a data-driven parameter learing method called Gaussian process regression (GPR) to predict optimal parameters. Numeri

Weichao Zhou, Wenchao Li

Many imitation learning (IL) algorithms use inverse reinforcement learning (IRL) to infer a reward function that aligns with the demonstration. However, the inferred reward functions often fail to capture the underlying task objectives. In this paper, we propose a novel framework for IRL-based IL that prioritizes task alignment over conventional data alignme

Harvey L. Devereux, Margaret-Ann Withington, Cillian Cockrell, Kostya Trachenko

We perform molecular dynamics simulations of molten Lithium Fluoride using the MACE-MP-0 (small) machine learnt interatomic potential and the classical Buckingham and Born-Huggins-Mayer potentials. We find that the MACE-MP-0, out-of-the-box, is able to accurately reproduce the experimental viscosity across the liquid state. Whilst the previous predicted visc

Wentao Zhang, Jingyuan Wang, Yifan Yang, Leong Hou U

Electronic maps consist of diverse entities, such as points of interest (POIs), road networks, and land parcels, playing a vital role in applications like ITS and LBS. Map entity representation learning (MapRL) generates versatile and reusable data representations, providing essential tools for efficiently managing and utilizing map entity data. Despite the

Zheng Chen, Buhui Yao

Goal hijacking is a type of adversarial attack on Large Language Models (LLMs) where the objective is to manipulate the model into producing a specific, predetermined output, regardless of the user's original input. In goal hijacking, an attacker typically appends a carefully crafted malicious suffix to the user's prompt, which coerces the model into ignorin

Soichiro Kumano, Hiroshi Kera, Toshihiko Yamasaki

Adversarial examples have raised several open questions, such as why they can deceive classifiers and transfer between different models. A prevailing hypothesis to explain these phenomena suggests that adversarial perturbations appear as random noise but contain class-specific features. This hypothesis is supported by the success of perturbation learning, wh

Mathilde Caron, Alireza Fathi, Cordelia Schmid, Ahmet Iscen

Web-scale visual entity recognition, the task of associating images with their corresponding entities within vast knowledge bases like Wikipedia, presents significant challenges due to the lack of clean, large-scale training data. In this paper, we propose a novel methodology to curate such a dataset, leveraging a multimodal large language model (LLM) for la

Liulin Li, Bingyi Liu, Zhongyi Guo

Recently, acoustic communication employing Orbital Angular Momentum (OAM) opens another avenue for efficient data transmission in aquatic environments. Current topological charge (TC) detection of OAM beams relies on the orthogonality among different-order OAM beams. Such strategy requires data collection from the entire acoustic field, which inevitably redu

Wenfeng Huang, Xinyun Liang, Jie Zhao, Zeliang Wu

Precise phase measurements by interferometers are crucial in science for detecting subtle changes, such as gravitational waves. However, phase sensitivity is typically limited by the standard quantum limit (SQL) with uncorrelated particles N. This limit can be surpassed using quantum correlations, but achieving high-quality correlations in large systems is c

Lingfeng Li, Jinniu Hu, Ying Zhang

The one-dimensional Kronig-Penney potential in the Schr\"{o}dinger equation, a standard periodic potential in quantum mechanics textbooks known for generating band structures, is solved by using the finite difference method with periodic boundary conditions. This method significantly improves the eigenvalue accuracy compared to existing approaches such as th

Junsoo Oh, Chulhee Yun

Patch-level data augmentation techniques such as Cutout and CutMix have demonstrated significant efficacy in enhancing the performance of vision tasks. However, a comprehensive theoretical understanding of these methods remains elusive. In this paper, we study two-layer neural networks trained using three distinct methods: vanilla training without augmentati

Van Thuy Hoang, Justin Widjaja, Y. Long Qiang, Maxwell Liu

The idea of fractional derivatives has a long history that dates back centuries. Apart from their intriguing mathematical properties, fractional derivatives have been studied widely in physics, for example in quantum mechanics and generally in systems with nonlocal temporal or spatial interactions. However, systematic experiments have been rare due to challe

Ivan Toftul, Sebastian Golat, Francisco J. Rodríguez-Fortuño, Franco Nori

The mechanical action of various kinds of waves has been recognized for several centuries. The first tide of scientific interest in wave-induced forces and torques emerged at the turn of the 20th century, with the development of wave theories and the concepts of wave momentum and angular momentum. A second surge occurred in the past several decades, driven b

Hassan Nagib, Ivan Marusic

Recently, Nagib et al (2024} utilized indicator functions of profiles of the streamwise normal stress to reveal the ranges of validity, in wall distance and Reynolds number, for each of two proposed models in DNS of channel and pipe flows. A method more suited to experimental data is proposed here, as establishing accurate indicator functions is a challenge.

David Koeplinger, Darshan Gandhi, Pushkar Nandkar, Nathan Sheeley

Token generation speed is critical to power the next wave of AI inference applications. GPUs significantly underperform during token generation due to synchronization overheads at kernel boundaries, utilizing only 21% of their peak memory bandwidth. While recent dataflow architectures mitigate these overheads by enabling aggressive fusion of decoder layers i

Alistair White, Anna Büttner, Maximilian Gelbrecht, Valentin Duruisseaux

Neural differential equations offer a powerful approach for learning dynamics from data. However, they do not impose known constraints that should be obeyed by the learned model. It is well-known that enforcing constraints in surrogate models can enhance their generalizability and numerical stability. In this paper, we introduce projected neural differential

Yang Xiao, Huai-Ke Guo, Jiahang Hu, Jin Min Yang

We derive a compact analytical expression for the growth factor $\Upsilon$, which characterizes how the gravitational wave spectrum sourced by sound waves evolves in a universe with a generic expansion history. Assuming the dominant energy density scales as $\rho \propto a^{-3(1+w)}$, we obtain $\Upsilon =\frac{2[1-y^{3(w-1)/2}]}{3(1-w)}$, where $y = a(t)/a(

Xue Li, Song Wang, Henggeng Han, Jifeng Liu

The ongoing discovery of exoplanets has sparked significant interest in finding suitable worlds that could potentially support life. Stellar ultraviolet (UV; 100-3000 \AA) radiation may play a crucial role in determining the habitability of their planets. In this paper, we conducted a detailed analysis of the UV photometry of over 2700 host stars with confir

Guram Bezhanishvili, Ramon Jansana

We develop a new duality for distributive and implicative meet semi-lattices. For distributive meet semi-lattices our duality generalizes Priestley's duality for distributive lattices and provides an improvement of Celani's duality. Our generalized Priestley spaces are similar to the ones constructed by Hansoul. Thus, one can view our duality for distributiv

Fan Nie, Jiangqun Ni, Jian Zhang, Bin Zhang

With the advancement of deepfake generation techniques, the importance of deepfake detection in protecting multimedia content integrity has become increasingly obvious. Recently, temporal inconsistency clues have been explored to improve the generalizability of deepfake video detection. According to our observation, the temporal artifacts of forged videos in

Shikang Yu, Tao Feng, Hengrui Liu

Let $a$, $b$ and $c$ be positive integers. Let $(G,+)$ be a finite abelian group of order $abc$. A $G$-magic rectangle set MRS$_G(a,b;c)$ is a collection of $c$ arrays of size $a\times b$ whose entries are elements of a group $G$, each appearing exactly once, such that the sum of each row in every array equals a constant $\gamma\in G$ and the sum of each col

Mingcong Han, Weihang Shen, Guanwen Peng, Rong Chen

We discovered that a GPU kernel can have both idempotent and non-idempotent instances depending on the input. These kernels, called conditionally-idempotent, are prevalent in real-world GPU applications (490 out of 547 from six applications). Consequently, prior work that classifies GPU kernels as either idempotent or non-idempotent can severely compromise t

Minghui Chen, Meirui Jiang, Xin Zhang, Qi Dou

Federated learning (FL) is a learning paradigm that enables collaborative training of models using decentralized data. Recently, the utilization of pre-trained weight initialization in FL has been demonstrated to effectively improve model performance. However, the evolving complexity of current pre-trained models, characterized by a substantial increase in p

Ziyue Yang, Chen Li, Yuqia Ran, Yongzhuo Li

Photonic linear operation is a promising approach to handle the extensive vector multiplications in artificial intelligence techniques due to the natural bosonic parallelism and high-speed information transmission of photonics. Although it is believed that maximizing the interaction of the light beams is necessary to fully utilize the parallelism and tremend

Dongwoo Lee, Joonkyu Park, Kyoung Mu Lee

To train a deblurring network, an appropriate dataset with paired blurry and sharp images is essential. Existing datasets collect blurry images either synthetically by aggregating consecutive sharp frames or using sophisticated camera systems to capture real blur. However, these methods offer limited diversity in blur types (blur trajectories) or require ext

Sadif Ahmed, Md Nafiu Rahman, Zahin Wahab, Gias Uddin

In the digital era, accidental exposure of sensitive information such as API keys, tokens, and credentials is a growing security threat. While most prior work focuses on detecting secrets in source code, leakage in software issue reports remains largely unexplored. This study fills that gap through a large-scale analysis and a practical detection pipeline fo

Iñigo Parra

This study investigates the impact of morphological typology on tokenization and language modeling performance. We focus on languages with synthetic and analytical morphological structures and examine their productivity when tokenized using the byte-pair encoding (BPE) algorithm. We compare the performance of models trained with similar amounts of data in di