October 2024 arXiv papers — page 104

Natsuko Yamaguchi, Kareem El-Badry, Nicholas M. Sorabella

Five self-lensing binaries (SLBs) have been discovered with Kepler light curves. They contain white dwarfs (WDs) in AU-scale orbits that gravitationally lens solar-type companions. Forming SLBs likely requires common envelope evolution when the WD progenitor is an AGB star and has a weakly bound envelope. No SLBs have yet been discovered with data from the T

Anuprita V. Kulkarni, Vatsana Tiwari, Auditya Sharma, Ankur Raina

Quantum simulation in its current state faces experimental overhead in terms of physical space and cooling. We propose boson sampling as an alternative compact synthetic platform performing at room temperature. Identifying the capability of estimating matrix permanents, we explore the applicability of boson sampling for tackling the dynamics of quantum syste

Santiago Cifuentes, Samson Wang, Thais L. Silva, Mario Berta

We investigate the dividing line between classical and quantum computational power in estimating properties of matrix functions. More precisely, we study the computational complexity of two primitive problems: given a function $f$ and a Hermitian matrix $A$, compute a matrix element of $f(A)$ or compute a local measurement on $f(A)|0\rangle^{\otimes n}$, wit

Carlos Miró, Miguel Escudero, Miguel Nebot

CP violation in neutral $B$ meson oscillations is an experimental observable that could be directly related to the baryon asymmetry of the Universe through the $B$-Mesogenesis mechanism. As this phenomenon is highly suppressed in the Standard Model, it could also be a sensitive probe for many new physics scenarios that modify neutral meson mixing. Motivated

Richard Dyer, Christopher J. Moore

Quasinormal modes (QNMs) are usually characterized by their time dependence; oscillations at specific frequencies predicted by black hole (BH) perturbation theory. QNMs are routinely identified in the ringdown of numerical relativity waveforms, are widely used in waveform modeling, and underpin key tests of general relativity and of the nature of compact obj

Francesco Perciavalle, Davide Rossini, Juan Polo, Luigi Amico

We explore the energy content of superpositions of current states. Specifically, we focus on the maximum energy that can be extracted from them through local unitary transformations. The figure of merit we employ is the local ergotropy. We perform a complete analysis in the whole range of the system's parameters. This way, we prove that superpositions of two

Alfredo Luminari, Enrico Piconcelli, Francesco Tombesi, Fabrizio Nicastro

Fast, massive winds are ubiquitously observed in the UV and X-ray spectra of Active Galactic Nuclei (AGN) and other accreting sources. Theoretical and observational evidences suggest they are launched at accretion disc scales, carrying significant mass and angular momentum. Thanks to such high energy output, they may play an important role in transferring th

Alexandros Ziampras, Cornelis P. Dullemond, Tilman Birnstiel, Myriam Benisty

Numerous protoplanetary disks exhibit shadows in scattered light observations. These shadows are typically cast by misaligned inner disks and are associated with observable structures in the outer disk such as bright arcs and spirals. Investigating the dynamics of the shadowed outer disk is therefore essential in understanding the formation and evolution of

Hayden Lee

We present a new method for evaluating tensor integrals in the large-scale structure. Decomposing a $\Lambda$CDM-like universe into a finite sum of scaling universes using the FFTLog, we can recast loop integrals for biased tracers in the large-scale structure as certain tensor integrals in quantum field theory. While rotational symmetry is spontaneously bro

Yijing Liu, Christopher Yang, Gabriel Gaertner, John Huckabee

Light-induced phenomena in materials can exhibit exotic behavior that extends beyond equilibrium properties, offering new avenues for understanding and controlling electronic phases. So far, non-equilibrium phenomena in solids have been predominantly explored using femtosecond laser pulses, which generate transient, ultra-fast dynamics. Here, we investigate

Valentina Capalbo, Marco De Petris, Antonio Ferragamo, Weiguang Cui

We analyse the maps of the Sunyaev-Zel'dovich (SZ) signal of local galaxy clusters ($z<0.1$) observed by the $Planck$ satellite in order to classify their dynamical state through morphological features. To study the morphology of the cluster maps, we apply a method recently employed on mock SZ images generated from hydrodynamical simulated galaxy clusters in

Ye Li, Wenzhao Zheng, Xiaonan Huang, Kurt Keutzer

Vision-centric autonomous driving has demonstrated excellent performance with economical sensors. As the fundamental step, 3D perception aims to infer 3D information from 2D images based on 3D-2D projection. This makes driving perception models susceptible to sensor configuration (e.g., camera intrinsics and extrinsics) variations. However, generalizing acro

Lijie Fan, Tianhong Li, Siyang Qin, Yuanzhen Li

Scaling up autoregressive models in vision has not proven as beneficial as in large language models. In this work, we investigate this scaling problem in the context of text-to-image generation, focusing on two critical factors: whether models use discrete or continuous tokens, and whether tokens are generated in a random or fixed raster order using BERT- or

Haofei Xu, Songyou Peng, Fangjinhua Wang, Hermann Blum

Gaussian splatting and single-view depth estimation are typically studied in isolation. In this paper, we present DepthSplat to connect Gaussian splatting and depth estimation and study their interactions. More specifically, we first contribute a robust multi-view depth model by leveraging pre-trained monocular depth features, leading to high-quality feed-fo

Rongyao Fang, Chengqi Duan, Kun Wang, Hao Li

Recent advancements in multimodal foundation models have yielded significant progress in vision-language understanding. Initial attempts have also explored the potential of multimodal large language models (MLLMs) for visual content generation. However, existing works have insufficiently addressed the varying granularity demands of different image generation

Runsen Xu, Zhiwei Huang, Tai Wang, Yilun Chen

3D visual grounding is crucial for robots, requiring integration of natural language and 3D scene understanding. Traditional methods depending on supervised learning with 3D point clouds are limited by scarce datasets. Recently zero-shot methods leveraging LLMs have been proposed to address the data issue. While effective, these methods only use object-centr

Yaxin Luo, Gen Luo, Jiayi Ji, Yiyi Zhou

Despite the significant progress in multimodal large language models (MLLMs), their high computational cost remains a barrier to real-world deployment. Inspired by the mixture of depths (MoDs) in natural language processing, we aim to address this limitation from the perspective of ``activated tokens''. Our key insight is that if most tokens are redundant fo

Tu Hong, Xiao Yan Xu

Non-Fermi liquids are an important topic in condensed matter physics, as their characteristics challenge the framework of traditional Fermi liquid theory and reveal the complex behavior of electrons in strongly interacting systems. Both the experimentally observed smeared region and the theoretically predicted marginal Fermi liquid suggest that spatial disor

Guhao Feng, Kai Yang, Yuntian Gu, Xinyue Ai

Despite the remarkable success of Transformer-based large language models (LLMs) across various domains, understanding and enhancing their mathematical capabilities remains a significant challenge. In this paper, we conduct a rigorous theoretical analysis of LLMs' mathematical abilities, with a specific focus on their arithmetic performances. We identify num

Cristina Cirstoiu

What makes a class of quantum circuits efficiently classically simulable on average? I present a framework that applies harmonic analysis of groups to circuits with a structure encoded by group parameters. Expanding the circuits in a suitable truncated multi-path operator basis gives algorithms to evaluate the Fourier coefficients of output distributions or

Runzhe Wu, Yiding Chen, Gokul Swamy, Kianté Brantley

Adversarial Imitation Learning is traditionally framed as a two-player zero-sum game between a learner and an adversarially chosen cost function, and can therefore be thought of as the sequential generalization of a Generative Adversarial Network (GAN). However, in recent years, diffusion models have emerged as a non-adversarial alternative to GANs that mere

Chenhao Zhang, Xi Feng, Yuelin Bai, Xinrun Du

As the capabilities of Multimodal Large Language Models (MLLMs) continue to improve, the need for higher-order capability evaluation of MLLMs is increasing. However, there is a lack of work evaluating MLLM for higher-order perception and understanding of Chinese visual content. To fill the gap, we introduce the **C**hinese **I**mage **I**mplication understan

Yifeng Wang, Xueying Zhan, Siyu Huang

As deep learning continues to evolve, the need for data efficiency becomes increasingly important. Considering labeling large datasets is both time-consuming and expensive, active learning (AL) provides a promising solution to this challenge by iteratively selecting the most informative subsets of examples to train deep neural networks, thereby reducing the

Zizhao Chen, Mustafa Omer Gul, Yiwei Chen, Gloria Geng

Multi-turn interactions between large language models (LLMs) and users naturally include implicit feedback signals. If an LLM responds in an unexpected way to an instruction, the user is likely to signal it by rephrasing the request, expressing frustration, or pivoting to an alternative task. Such signals are task-independent and occupy a relatively constrai

Ruoshi Liu, Alper Canberk, Shuran Song, Carl Vondrick

Vision foundation models trained on massive amounts of visual data have shown unprecedented reasoning and planning skills in open-world settings. A key challenge in applying them to robotic tasks is the modality gap between visual data and action data. We introduce differentiable robot rendering, a method allowing the visual appearance of a robot body to be

Bruno Mlodozeniec, Runa Eschenhagen, Juhan Bae, Alexander Immer

Diffusion models have led to significant advancements in generative modelling. Yet their widespread adoption poses challenges regarding data attribution and interpretability. In this paper, we aim to help address such challenges in diffusion models by developing an influence functions framework. Influence function-based data attribution methods approximate h

Florian Hübler, Ilyas Fatkhullin, Niao He

Recent empirical evidence indicates that many machine learning applications involve heavy-tailed gradient noise, which challenges the standard assumptions of bounded variance in stochastic optimization. Gradient clipping has emerged as a popular tool to handle this heavy-tailed noise, as it achieves good performance in this setting both theoretically and pra

Chengyue Wu, Xiaokang Chen, Zhiyu Wu, Yiyang Ma

In this paper, we introduce Janus, an autoregressive framework that unifies multimodal understanding and generation. Prior research often relies on a single visual encoder for both tasks, such as Chameleon. However, due to the differing levels of information granularity required by multimodal understanding and generation, this approach can lead to suboptimal

Ariel Slepyan, Dian Li, Hongjun Cai, Ryan McGovern

Robots require full-body, high-resolution tactile sensing to operate safely in unstructured environments, enabling reflexive responses and closed-loop control. However, the pixel counts needed for dense, large-area coverage limit readout rates of most tactile arrays to <100 Hz, hindering their use in high-speed tasks. We present Adaptive Compressive Tactile

Eric J. Ching, Ryan F. Johnson

Standing detonation engines are a promising detonation-based propulsion technology. The most commonly studied standing detonation configuration involves a straight-sided wedge that induces an oblique detonation wave. A recently introduced standing-detonation-engine concept entails a curved ramp that leads to formation of a curved detonation wave. The continu

Kriti Baweja, David J. Luitz, Samuel J. Garratt

We show how the effects of large numbers of measurements on many-body quantum ground and thermal states can be studied using Quantum Monte Carlo (QMC). Density matrices generated by measurement in this setting feature products of many local non-unitary operators, and by expanding these density matrices as sums over operator strings we arrive at a generalized

Nicolás Parra, Spyros Sypsas, Gonzalo A. Palma, Cristóbal Zenteno

Particle production in de Sitter spacetime arises from the exponential expansion of space, rendering the Bunch-Davies vacuum perceived as a particle-containing state by late-time observers. For states defined as eigenstates of both momentum and the Hamiltonian, the Bunch-Davies vacuum exhibits a constant particle density per physical momentum. We explore par

Yansong Peng, Hebei Li, Peixi Wu, Yueyi Zhang

We introduce D-FINE, a powerful real-time object detector that achieves outstanding localization precision by redefining the bounding box regression task in DETR models. D-FINE comprises two key components: Fine-grained Distribution Refinement (FDR) and Global Optimal Localization Self-Distillation (GO-LSD). FDR transforms the regression process from predict

Qiaoyu Tang, Le Yu, Bowen Yu, Hongyu Lin

Post-training has emerged as a crucial paradigm for adapting large-scale pre-trained models to various tasks, whose effects are fully reflected by delta parameters (i.e., the disparity between post-trained and pre-trained parameters). While numerous studies have explored delta parameter properties via operations like pruning, quantization, low-rank approxima

Parikshit Chalise, Antwan Clark, Edinah K. Gnang

We prove a conjecture of Gyárfás (1976), which asserts that any family of trees $T_1, \dots, T_{n}$ where each $T_k$ has $k$ vertices packs into $K_n$. We do so by translating the decomposition problem into a labeling problem, namely complete labeling. Our proof employs the polynomial method using a functional reformulation of the conjecture.

Gonçalo Paulo, Alex Mallen, Caden Juang, Nora Belrose

While the activations of neurons in deep neural networks usually do not have a simple human-understandable interpretation, sparse autoencoders (SAEs) can be used to transform these activations into a higher-dimensional latent space which may be more easily interpretable. However, these SAEs can have millions of distinct latent features, making it infeasible

Tan Dat Nguyen, Ji-Hoon Kim, Jeongsoo Choi, Shukjae Choi

The goal of this paper is to accelerate codec-based speech synthesis systems with minimum sacrifice to speech quality. We propose an enhanced inference method that allows for flexible trade-offs between speed and quality during inference without requiring additional training. Our core idea is to predict multiple tokens per inference step of the AR module usi

Darja Nonaca, Christoph Studer

Massive multiuser (MU) multiple-input multiple-output (MIMO) enables concurrent transmission of multiple users to a multi-antenna basestation (BS). To detect the users' data using linear equalization, the BS must perform preprocessing, which requires, among other tasks, the inversion of a matrix whose dimension equals the number of user data streams. Explici

Chen Bo Calvin Zhang, Zhang-Wei Hong, Aldo Pacchiano, Pulkit Agrawal

Reward shaping is critical in reinforcement learning (RL), particularly for complex tasks where sparse rewards can hinder learning. However, choosing effective shaping rewards from a set of reward functions in a computationally efficient manner remains an open challenge. We propose Online Reward Selection and Policy Optimization (ORSO), a novel approach that

André Platzer, Long Qian

This article proves the completeness of an axiomatization for initial value problems (IVPs) with compact initial conditions and compact time horizons for bounded open safety, open liveness and existence properties. Completeness systematically reduces the proofs of these properties to a complete axiomatization for differential equation invariants. This result

Tianyu Guo, Druv Pai, Yu Bai, Jiantao Jiao

Practitioners have consistently observed three puzzling phenomena in transformer-based large language models (LLMs): attention sinks, value-state drains, and residual-state peaks, collectively referred to as extreme-token phenomena. These phenomena are characterized by certain so-called "sink tokens" receiving disproportionately high attention weights, exhib

Zsuzsanna Baran

We consider random walks on $\Z^8$ indexed by the infinite invariant tree, which consists of an infinite spine and finite random trees attached to it on both sides. We establish the precise order of the non-intersection probability between one walk indexed by one side of the tree, and an independent one indexed by both sides of an independent tree. This is a

Malte Esders, Thomas Schnake, Jonas Lederer, Adil Kabylda

While machine learning (ML) models have been able to achieve unprecedented accuracies across various prediction tasks in quantum chemistry, it is now apparent that accuracy on a test set alone is not a guarantee for robust chemical modeling such as stable molecular dynamics (MD). To go beyond accuracy, we use explainable artificial intelligence (XAI) techniq

Jingwei Ma, Erika Lu, Roni Paiss, Shiran Zada

Panoramic image stitching provides a unified, wide-angle view of a scene that extends beyond the camera's field of view. Stitching frames of a panning video into a panoramic photograph is a well-understood problem for stationary scenes, but when objects are moving, a still panorama cannot capture the scene. We present a method for synthesizing a panoramic vi

Kajetan Schweighofer, Adrian Arnaiz-Rodriguez, Sepp Hochreiter, Nuria Oliver

Ensembles of Deep Neural Networks, Deep Ensembles, are widely used as a simple way to boost predictive performance. However, their impact on algorithmic fairness is not well understood yet. Algorithmic fairness examines how a model's performance varies across socially relevant groups defined by protected attributes such as age, gender, or race. In this work,

Yujie Wei, Shiwei Zhang, Hangjie Yuan, Xiang Wang

Recent advances in customized video generation have enabled users to create videos tailored to both specific subjects and motion trajectories. However, existing methods often require complicated test-time fine-tuning and struggle with balancing subject learning and motion control, limiting their real-world applications. In this paper, we present DreamVideo-2

Magnus Sesodia, Shinichi Sunami, En Chang, Erik Rydow

Quantum vortices play a crucial role in both equilibrium and dynamical phenomena in two-dimensional (2D) superfluid systems. Experimental detection of these excitations in 2D ultracold atomic gases typically involves examining density depletions in absorption images, however the presence of a significant phononic background renders the problem challenging, b

Hui Yuan, Yifan Zeng, Yue Wu, Huazheng Wang

Reinforcement Learning from Human Feedback (RLHF) has become the predominant approach for language model (LM) alignment. At its core, RLHF uses a margin-based loss for preference optimization, specifying ideal LM behavior only by the difference between preferred and dispreferred responses. In this paper, we identify a common pitfall of margin-based methods -

Nicholas Bohlsen

Three novel applications of computational topology in the field of fusion science are developed. A procedure for the automatic classification of the orbits of magnetic field lines into topologically distinct classes using Vietoris-Rips persistent homology is presented and tested for a toy model of a perturbed tokamak. A method for estimating the distribution

Sebastián Rodríguez-Martínez, Giancarlo Troni

MEMS Attitude Heading Reference Systems are widely employed to determine a system's attitude, but sensor measurement biases limit their accuracy. This paper introduces a novel factor graph-based method called MAgnetometer and GYroscope Calibration (MAGYC). MAGYC leverages three-axis angular rate measurements from an angular rate gyroscope to enhance calibrat

Mazda Moayeri, Vidhisha Balachandran, Varun Chandrasekaran, Safoora Yousefi

With models getting stronger, evaluations have grown more complex, testing multiple skills in one benchmark and even in the same instance at once. However, skill-wise performance is obscured when inspecting aggregate accuracy, under-utilizing the rich signal modern benchmarks contain. We propose an automatic approach to recover the underlying skills relevant

Ke Yang, Yao Liu, Sapana Chaudhary, Rasool Fakoor

Autonomy via agents using large language models (LLMs) for personalized, standardized tasks boosts human efficiency. Automating web tasks (like booking hotels within a budget) is increasingly sought after. Fulfilling practical needs, the web agent also serves as an important proof-of-concept example for various agent grounding scenarios, with its success pro

Junpeng Liu, Tianyue Ou, Yifan Song, Yuxiao Qu

Text-rich visual understanding-the ability to process environments where dense textual content is integrated with visuals-is crucial for multimodal large language models (MLLMs) to interact effectively with structured environments. To enhance this capability, we propose synthesizing general multimodal instructions from webpage UIs using text-based large lang

Xiaodan Xing, Junzhi Ning, Yang Nan, Guang Yang

Deep generative models have significantly advanced medical imaging analysis by enhancing dataset size and quality. Beyond mere data augmentation, our research in this paper highlights an additional, significant capacity of deep generative models: their ability to reveal and demonstrate patterns in medical images. We employ a generative structure with hybrid

Clément Playout, Renaud Duval, Marie Carole Boucher, Farida Cheriet

The diagnosis of diabetic retinopathy, which relies on fundus images, faces challenges in achieving transparency and interpretability when using a global classification approach. However, segmentation-based databases are significantly more expensive to acquire and combining them is often problematic. This paper introduces a novel method, termed adversarial s

Takeru Miyato, Sindy Löwe, Andreas Geiger, Max Welling

It has long been known in both neuroscience and AI that ``binding'' between neurons leads to a form of competitive learning where representations are compressed in order to represent more abstract concepts in deeper layers of the network. More recently, it was also hypothesized that dynamic (spatiotemporal) representations play an important role in both neur

Rongzhi Gao, ChiYung Yam, Jianjun Mao, Shuguang Chen

Long-range interactions are essential determinants of chemical system behaviour across diverse environments. We present a foundation framework that integrates explicit polarizable long-range physics with an equivariant graph neural network potential. It employs a physically motivated polarizable charge equilibration scheme that directly optimizes electrostat

Emily Adlam

Recent developments in foundations of physics have given rise to a class of views suggesting that physically meaningful descriptions must always be relativized to a physical perspective. In this article I distinguish between strong physical perspectivalism, which maintains that all facts must be relativized to a perspective, and moderate physical perspectiva

Elena Cordero, Gianluca Giacchi, Eugenia Malinnikova

Hardy's uncertainty principle is a classical result in harmonic analysis, stating that a function in $L^2(\mathbb{R}^d)$ and its Fourier transform cannot both decay arbitrarily fast at infinity. In this paper, we extend this principle to the propagators of Schr\"odinger equations with quadratic Hamiltonians, known in the literature as metaplectic operators.

Jean-Pierre Sleiman, Mayank Mittal, Marco Hutter

Reinforcement learning (RL) often necessitates a meticulous Markov Decision Process (MDP) design tailored to each task. This work aims to address this challenge by proposing a systematic approach to behavior synthesis and control for multi-contact loco-manipulation tasks, such as navigating spring-loaded doors and manipulating heavy dishwashers. We define a

Mitsuhiko Nakamoto, Oier Mees, Aviral Kumar, Sergey Levine

Large, general-purpose robotic policies trained on diverse demonstration datasets have been shown to be remarkably effective both for controlling a variety of robots in a range of different scenes, and for acquiring broad repertoires of manipulation skills. However, the data that such policies are trained on is generally of mixed quality -- not only are huma

Arinjoy De, Alessio Lerose, De Luo, Federica M. Surace

Spontaneous particle-pair formation is a fundamental phenomenon in nature. It can, for example, appear when the potential energy between two particles increases with separation, as if they were connected by a tense string. Beyond a critical separation, new particle pairs can form, causing the string to break. String-breaking dynamics in quantum chromodynamic

Ning Bao, Gun Suer

The advent of fault-tolerant quantum computers marks a significant milestone, yet the development of practical quantum algorithms remains a critical challenge. Effective quantum algorithms are essential for leveraging the power of quantum computers, and their design is often non-intuitive. This paper addresses the issue of generating novel quantum primitives

Torben Fattler, Martin Grothaus, Nathalie Steil

The starting point is a gradient Dirichlet form with respect to $\varrho\lambda^d$ on the space $L^2({\mathbb{R}}^d, \varrho\mu)$. Here $\lambda^d$ is the Lebesgue measure on ${\mathbb R}^d$, $\varrho$ a strictly positive density and $\mu$ puts weight on a set $A\subset {\mathbb R}^d$ with Lebesgue measure zero. We show that the Dirichlet form admits an asso

Sepehr Sadoughi, Nikolay Yakovets, George Fletcher

The ISO standard Property Graph model has become increasingly popular for representing complex, interconnected data. However, it lacks native support for querying metadata and reification, which limits its abilities to deal with the demands of modern applications. We introduce the vision of Meta-Property Graphs, a backwards compatible extension of the proper

Mohamed Nomeir, Pasan Dissanayake, Shreya Meel, Sanghamitra Dutta

Transparency and explainability are two extremely important aspects to be considered when employing black-box machine learning models in high-stake applications. Providing counterfactual explanations is one way of fulfilling this requirement. However, this also poses a threat to the privacy of both the institution that is providing the explanation as well as

K. Senthil Kumar, K. M. B. Abdullah Safwan

Object Detection is related to Computer Vision. Object detection enables detecting instances of objects in images and videos. Due to its increased utilization in surveillance, tracking system used in security and many others applications have propelled researchers to continuously derive more efficient and competitive algorithms. However, problems emerges whi

Matteo Gallet, Georg Grasegger, Jan Legerský, Josef Schicho

Steffen's polyhedron was believed to have the least number of vertices among polyhedra that can flex without self-intersections. Maksimov clarified that the pentagonal bipyramid with one face subdivided into three is the only polyhedron with fewer vertices for which the existence of a self-intersection-free flex was open. Since subdividing a face into three

Amirhosein Alizadeh, Seyed Fariborz Zarei, Mohammadhadi Shateri

This paper proposes an AI-based scheme for islanding detection in active distribution networks. By reviewing existing studies, it is clear that there are several gaps in the field to ensure reliable islanding detection, including (i) model complexity and stability concerns, (ii) limited accuracy under noisy conditions, and (iii) limited applicability to syst

Eric J. Ching, Ryan F. Johnson

This paper presents a conservative discontinuous Galerkin method for the simulation of supercritical and transcritical real-fluid flows without phase separation. A well-known issue associated with the use of fully conservative schemes is the generation of spurious pressure oscillations at contact interfaces, which are exacerbated when a cubic equation of sta

Xingang Guo, Darioush Keivan, Usman Syed, Lianhui Qin

Control system design is a crucial aspect of modern engineering with far-reaching applications across diverse sectors including aerospace, automotive systems, power grids, and robotics. Despite advances made by Large Language Models (LLMs) in various domains, their application in control system design remains limited due to the complexity and specificity of

Paul M. N. Feehan, Thomas G. Leness

This work is a sequel to our previous monograph arXiv:2010.15789 (to appear in AMS Memoirs), where we initiated our program to prove that the Bogomolov-Miyaoka-Yau inequality holds for closed, symplectic four-manifolds and, more generally, for closed, smooth four-manifolds with a Seiberg-Witten basic class. This inequality was first proved for compact, compl

Ruibo Chen, Yihan Wu, Junfeng Guo, Heng Huang

Watermarking techniques offer a promising way to identify machine-generated content via embedding covert information into the contents generated from language models (LMs). However, the robustness of the watermarking schemes has not been well explored. In this paper, we present De-mark, an advanced framework designed to remove n-gram-based watermarks effecti

Junhao Gu, Peng-Tao Jiang, Hao Zhang, Mi Zhou

Recent methods exploit the powerful text-to-image (T2I) diffusion models for real-world image super-resolution (Real-ISR) and achieve impressive results compared to previous models. However, we observe two kinds of inconsistencies in diffusion-based methods which hinder existing models from fully exploiting diffusion priors. The first is the semantic inconsi

Jeongjae Lee, Songnam Hong

We study the channel estimation problem for a reconfigurable intelligent surface (RIS)-assisted millimeter-wave (mmWave) multi-user multiple-input multiple-output (MU-MIMO) system. In particular, it is assumed that the channel between a RIS and a base station (BS) exhibits a near-field line-of-sight (LoS) channel, which is a dominant signal path in mmWave co

Ruibo Chen, Yihan Wu, Yanshuo Chen, Chenxi Liu

Statistical watermarking techniques are well-established for sequentially decoded language models (LMs). However, these techniques cannot be directly applied to order-agnostic LMs, as the tokens in order-agnostic LMs are not generated sequentially. In this work, we introduce Pattern-mark, a pattern-based watermarking framework specifically designed for order

Hongyu Zhao, Ming Li, Lichao Sun, Tianyi Zhou

Evaluating large language models (LLMs) is costly: it requires the generation and examination of LLM outputs on a large-scale benchmark of various tasks. This paper investigates how to efficiently reduce the tasks used to benchmark LLMs without affecting the evaluation quality. Our study reveals that task transferability and relevance provide critical inform

Gianluca Apriceno, Valentina Tamma, Tania Bailoni, Jacopo de Berardinis

The ability to reason with and integrate different sensory inputs is the foundation underpinning human intelligence and it is the reason for the growing interest in modelling multi-modal information within Knowledge Graphs. Multi-Modal Knowledge Graphs extend traditional Knowledge Graphs by associating an entity with its possible modal representations, inclu

Patrik Zavoral, Dušan Variš, Ondřej Bojar

The Transformer model has a tendency to overfit various aspects of the training data, such as the overall sequence length. We study elementary string edit functions using a defined set of error indicators to interpret the behaviour of the sequence-to-sequence Transformer. We show that generalization to shorter sequences is often possible, but confirm that lo

Nazanin Hoghooghi, Mikael Mazur, Nicolas Fontaine, Yifan Liu

The telecommunications industry's deployment of billions of kilometers of optical fiber has created a vast global network that can be exploited for additional applications such as environmental sensing, quantum networking and international clock comparisons. However, for reasons such as the unidirectionality of long-haul fiber links, telecom fiber networks c

Ernesto Arganda, Marcela Carena, Martín de los Rios, Andres D. Perez

The search for weakly interacting matter particles (WIMPs) is one of the main objectives of the High Luminosity Large Hadron Collider (HL-LHC). In this work we use Machine-Learning (ML) techniques to explore WIMP radiative decays into a Dark Matter (DM) candidate in a supersymmetric framework. The minimal supersymmetric WIMP sector includes the lightest neut

Limei Wang, Kaveh Hassani, Si Zhang, Dongqi Fu

Transformers serve as the backbone architectures of Foundational Models, where domain-specific tokenizers allow them to adapt to various domains. Graph Transformers (GTs) have recently emerged as leading models in geometric deep learning, outperforming Graph Neural Networks (GNNs) in various graph learning tasks. However, the development of tokenizers for gr

Ranajay Datta, Peter Virnau

We investigate with numerical simulations the influence of topology and stiffness on macroscopic rheological properties of polymer melts consisting of unknotted, knotted or concatenated rings. While melts of flexible, knotted oligomer rings tend to be significantly more viscous than their unknotted counterparts, differences vanish in a low shear rate scenari

Daniel E Widdowson, Vitaliy A Kurlin

With the advent of self-driving labs promising to synthesize large numbers of new materials, new automated tools are required for checking potential duplicates in existing structural databases before a material can be claimed as novel. To avoid duplication, we rigorously define the novelty metric of any periodic material as the smallest distance to its neare

Dmitri LaBelle, Yong-Jie Hu

Phase-pure synthesis has been a major challenge for metal oxynitrides due to their sensitivity to synthesis conditions and the limited understanding of the underlying thermodynamics. The beta-phase tantalum oxynitride (beta-TaON), a promising material for applications in photocatalysis and energy storage, is particularly difficult to synthesize in a reproduc

Da Long, Zhitong Xu, Guang Yang, Akil Narayan

Modern physics simulation often involves multiple functions of interests, and traditional numerical approaches are known to be complex and computationally costly. While machine learning-based surrogate models can offer significant cost reductions, most focus on a single task, such as forward prediction, and typically lack uncertainty quantification -- an ess

Jonas P. Pereira, Tulio Ottoni, Jaziel G. Coelho, Jorge A. Rueda

Rigid (Uniform) rotation is usually assumed when investigating the properties of mature neutron stars (NSs). Although it simplifies their description, it is an assumption because we cannot observe the NS's innermost parts. Here, we analyze the structure of NSs in the simple case of ''almost rigidity,'' where the innermost and outermost parts rotate with diff

Ilya Kaufman, Omri Azencot

Transformer models have consistently achieved remarkable results in various domains such as natural language processing and computer vision. However, despite ongoing research efforts to better understand these models, the field still lacks a comprehensive understanding. This is particularly true for deep time series forecasting methods, where analysis and un

Soumangsu Chakraborty, Akikazu Hashimoto, Horatiu Nastase

We revisit the problem of computing the determinant of Klein-Gordon operator $\Delta = -\nabla^2 + M^2$ on Euclideanized $AdS_3$ with the Euclideanized time coordinate compactified with period $\beta$, $H_3/Z$, by explicitly computing its eigenvalues and computing their product. Upon assuming that eigenfunctions are normalizable on $H_3/Z$, we found that the

Liang Xu, Shaoyang Hua, Zili Lin, Yifan Liu

In this paper, we tackle the problem of how to build and benchmark a large motion model (LMM). The ultimate goal of LMM is to serve as a foundation model for versatile motion-related tasks, e.g., human motion generation, with interpretability and generalizability. Though advanced, recent LMM-related works are still limited by small-scale motion data and cost

Tanja Kirchner, Matthias Göbel, Hans-Werner Hammer

We compute the neutron-neutron relative-energy distribution of the triton following the hard knockout of the proton in pionless effective field theory. This distribution can be used to study universality as well as to obtain information on the neutron-neutron interaction. Especially, one can infer the scattering length from fitting theory predictions for the

Michael J. Q. Zhang, W. Bradley Knox, Eunsol Choi

Large language models (LLMs) must often respond to highly ambiguous user requests. In such cases, the LLM's best response may be to ask a clarifying question to elicit more information. Existing LLMs often respond by presupposing a single interpretation of such ambiguous requests, frustrating users who intended a different interpretation. We speculate this i

Felix J Binder, James Chua, Tomek Korbak, Henry Sleight

Humans acquire knowledge by observing the external world, but also by introspection. Introspection gives a person privileged access to their current state of mind (e.g., thoughts and feelings) that is not accessible to external observers. Can LLMs introspect? We define introspection as acquiring knowledge that is not contained in or derived from training dat

Fengqi Liu, Hexiang Wang, Jingyu Gong, Ran Yi

Speech-driven gesture generation aims at synthesizing a gesture sequence synchronized with the input speech signal. Previous methods leverage neural networks to directly map a compact audio representation to the gesture sequence, ignoring the semantic association of different modalities and failing to deal with salient gestures. In this paper, we propose a n

Zekun Moore Wang, Shawn Wang, Kang Zhu, Jiaheng Liu

Alignment of large language models (LLMs) involves training models on preference-contrastive output pairs to adjust their responses according to human preferences. To obtain such contrastive pairs, traditional methods like RLHF and RLAIF rely on limited contrasting patterns, such as varying model variants or decoding temperatures. This singularity leads to t

Sindura Saraswathi, Christian Kümmerle

The Lightning Network is a peer-to-peer network designed to address Bitcoin's scalability challenges, facilitating rapid, cost-effective, and instantaneous transactions through bidirectional, blockchain-backed payment channels among network peers. Due to a source-based routing of payments, different pathfinding strategies are used in practice, trading off di

Idris Abdulmumin, Bashir Shehu Galadanci, Garba Aliyu, Shamsuddeen Hassan Muhammad

Monolingual data, being readily available in large quantities, has been used to upscale the scarcely available parallel data to train better models for automatic translation. Self-learning, where a model is made to learn from its output, is one approach to exploit such data. However, it has been shown that too much of this data can be detrimental to the perf

Xinyou Wang, Zaixiang Zheng, Fei Ye, Dongyu Xue

Proteins are essential macromolecules defined by their amino acid sequences, which determine their three-dimensional structures and, consequently, their functions in all living organisms. Therefore, generative protein modeling necessitates a multimodal approach to simultaneously model, understand, and generate both sequences and structures. However, existing

Yu Jia, Shuo Lin, Jian Zhou, Ya-jin Zhou

Elastic light-by-light(LbL) scattering, one of the most fascinating processes in the Standard Model(SM), has recently been observed in the ultraperipheral collisions(UPCs) of relativistic heavy ions in the Atlas and CMS experiments at the Large Hadron Collider LHC. Recognizing that the incident quasi-real photons in LbL scattering are strongly linearly polar

Or Ordentlich, Yury Polyanskiy

Recent work in machine learning community proposed multiple methods for performing lossy compression (quantization) of large matrices. This quantization is important for accelerating matrix multiplication (main component of large language models), which is often bottlenecked by the speed of loading these matrices from memory. Unlike classical vector quantiza