November 2023 arXiv papers — page 2

Siqi Wang, Chau Pham, Bryan A. Plummer

Noisy labels can impair model performance, making the study of learning with noisy labels an important topic. Two conventional approaches are noise modeling and noise detection. However, these two methods are typically studied independently, and there has been limited work on their collaboration. In this work, we explore the integration of these two approach

Federico Ettori, Thibaud Coupé, Timothy J. Sluckin, Ezio Puppin

We investigate the dynamic phase transition in two-dimensional Ising models whose equilibrium characteristics are influenced by either anisotropic interactions or quenched defects. The presence of anisotropy reduces the dynamical critical temperature, leading to the expected result that the critical temperature approaches zero in the full-anisotropy limit. W

C. Klein, F. Linares, D. Pilod, J. -C. Saut

We consider in this paper various theoretical and numerical issues on classical one dimensional models of internal waves with surface tension.They concern the Cauchy problem, including the long time dynamic, localized solitons or multisolitons, the soliton resolution property. We survey known results, present a few new ones together with open questions and c

Meera Hahn, Amit Raj, James M. Rehg

The challenging task of Vision-and-Language Navigation (VLN) requires embodied agents to follow natural language instructions to reach a goal location or object (e.g. `walk down the hallway and turn left at the piano'). For agents to complete this task successfully, they must be able to ground objects referenced into the instruction (e.g.`piano') into the vi

Tamar Meshveliani, Mark R. Lovell, Robert A. Crain, Joel Pfeffer

We study the statistics of dwarf galaxy populations as a function of environment in the cold dark matter (CDM) and warm dark matter (WDM) cosmogonies, using hydrodynamical simulations starting from initial conditions with matched phases but differing power spectra, and evolved with the EAGLE galaxy formation model. We measure the abundance of dwarf galaxies

Sadeep Jayasumana, Srikumar Ramalingam, Andreas Veit, Daniel Glasner

As with many machine learning problems, the progress of image generation methods hinges on good evaluation metrics. One of the most popular is the Frechet Inception Distance (FID). FID estimates the distance between a distribution of Inception-v3 features of real images, and those of images generated by the algorithm. We highlight important drawbacks of FID:

Silvia Paparini, Epifanio G. Virga

The elastic theory of chromonic liquid crystals is not completely established. We know, for example, that for anomalously low twist constants (needed for chromonics) the classical Oseen-Frank theory may entail paradoxical consequences when applied to describe the equilibrium shapes of droplets surrounded by an isotropic phase: contrary to experimental eviden

Steven DiSilvio, Anthony Ozerov, Leon Zhou

With the advent of power-meters allowing cyclists to precisely track their power outputs throughout the duration of a race, devising optimal power output strategies for races has become increasingly important in competitive cycling. To do so, the track, weather, and individual cyclist's abilities must all be considered. We propose differential equation model

Ihor Yatsuta, David F. Mross

The Kitaev honeycomb model supports gapless and gapped quantum spin liquid phases. Its exact solvability relies on extensively many locally conserved quantities. Any real-world manifestation of these phases would include imperfections in the form of disorder and interactions that break integrability. We show that the latter qualitatively alters the propertie

Valerio Lupi, J. Simon Kern, Philipp Schlatter

This study analyses the main characteristics of the fully developed laminar pulsatile flow in a toroidal pipe as the governing parameters vary. A novel computational technique is developed to obtain time-periodic solutions of the Navier$\unicode{x2013}$Stokes equations. They are computed as fixed points of the system in the frequency domain via a Newton$\uni

Le Xuan The Tai, Mattia Longobucco, Nguyen Viet Hung, Bartosz Paluba

We systematically present experimental and theoretical results for the dual-wavelength switching of 1560 nm, 75 fs signal pulses (SPs) driven by 1030 nm, 270 fs control pulses (CPs) in a dual-core fiber (DCF). We demonstrate a switching contrast of 31.9 dB, corresponding to a propagation distance of 14 mm, achieved by launching temporally synchronized SP-CP

Alena Pirutka

In this note we define a subgroup $H^i_{nr,\pi}$ of unramified cohomology group $H^i_{nr}$ of a fibration $\pi:X\to S$. This subgroup can be used efficiently in refined specialization arguments and allows to detect the failure of stable rationality of a variety specializing to $X$. We compute $H^2_{nr, \pi}$ systematically for many cubic surface bundles $\pi

Brian F. Healy, Michael W. Coughlin, Ashish A. Mahabal, Theophile Jegou du Laz

The classification of variable objects provides insight into a wide variety of astrophysics ranging from stellar interiors to galactic nuclei. The Zwicky Transient Facility (ZTF) provides time series observations that record the variability of more than a billion sources. The scale of these data necessitates automated approaches to make a thorough analysis.

Guido Dittrich, Luisa G. Cencha, Martin Steinhart, Ralf B. Wehrspohn

Polymers are known to wet nanopores with high surface energy through an atomically thin precursor film followed by slower capillary filling. We present here light interference spectroscopy using a nanoporous membrane-based chip that allows us to observe the dynamics of these phenomena in situ with sub-nanometer spatial and milli- to microsecond temporal reso

Luke Finnerty, Jerry W. Xuan, Yinzi Xin, Joshua Liberman

We present high-resolution $K$-band emission spectra of the quintessential hot Jupiter HD 189733 b from the Keck Planet Imager and Characterizer (KPIC). Using a Bayesian retrieval framework, we fit the dayside pressure-temperature profile, orbital kinematics, mass-mixing ratios of H$_2$O, CO, CH$_4$, NH$_3$, HCN, and H$_2$S, and the $\rm ^{13}CO/^{12}CO$ rat

Robert van Steenbergen, Wouter van Heeswijk, Martijn Mes

Humanitarian logistics operations face increasing difficulties due to rising demands for aid in disaster areas. This paper investigates the dynamic allocation of scarce relief supplies across multiple affected districts over time. It introduces a novel stochastic dynamic post-disaster inventory allocation problem with trucks and unmanned aerial vehicles deli

Robert Stein, Ashish Mahabal, Simeon Reusch, Matthew Graham

Optical surveys have become increasingly adept at identifying candidate Tidal Disruption Events (TDEs) in large numbers, but classifying these generally requires extensive spectroscopic resources. Here we present $\texttt{tdescore}$, a simple binary photometric classifier that is trained using a systematic census of $\sim$3000 nuclear transients from the Zwi

Demetre Kazaras, Antoine Song, Kai Xu

We show that any closed hyperbolic 3-manifold M admits a Riemannian metric with scalar curvature at least -6, but with volume entropy strictly larger than 2. In particular, this construction gives counterexamples to a conjecture of I. Agol, P. Storm and W. Thurston.

Matthew J. Colbrook

Dynamic Mode Decomposition (DMD) is a popular data-driven analysis technique used to decompose complex, nonlinear systems into a set of modes, revealing underlying patterns and dynamics through spectral analysis. This review presents a comprehensive and pedagogical examination of DMD, emphasizing the role of Koopman operators in transforming complex nonlinea

Rouzbeh Allahverdi, Chiara Arina, Marco Chianese, Michele Cicoli

We study the phenomenology of superheavy decaying dark matter with mass around $10^{10}$ GeV which can arise in the low-energy limit of string compactifications. Generic features of string theory setups (such as high scale supersymmetry breaking and epochs of early matter domination driven by string moduli) can accommodate superheavy dark matter with the cor

Manthos Karydas, Songyuan Li, Anastasios C. Petkou, Matthieu Vilatte

We show that a class of $L$-loop conformal ladder graphs correspond to twisted partition functions of free massive complex scalars in $d=2L+1$ dimensions. The graphs arise as four-point functions in certain two- and four-dimensional conformal fishnet models. The twisted thermal two-point function of the scalars is a generator of such conformal graphs for all

Hendra I. Nurdin

Quantum Markov models are employed ubiquitously in quantum physics and in quantum information theory due to their relative simplicity and analytical tractability. In particular, these models are known to give accurate approximations for a wide range of quantum optical and mesoscopic systems. However, in general, the validity of the Markov approximation entai

J. J. Blanco-Pillado, J. R. Espinosa, J. Huertas, K. Sousa

Bubbles of nothing (BoNs) describe the decay of spacetimes with compact dimensions and are thus of fundamental importance for many higher dimensional theories proposed beyond the Standard Model. BoNs admit a 4-dimensional description in terms of a singular Coleman-de Luccia (CdL) instanton involving the size modulus field, stabilized by some potential $V(\ph

Mircea Bejan, Campbell McLauchlan, Benjamin Béri

The classical simulation of highly-entangling quantum dynamics is conjectured to be generically hard. Thus, recently discovered measurement-induced transitions between highly entangling and low-entanglement dynamics are phase transitions in classical simulability. Here, we study simulability transitions beyond entanglement: noting that some highly-entangling

Itai Sfaradi, Assaf Horesh, Jesper Sollerman, Rob Fender

We present here broadband radio observations of the CSM interacting SN2019oys. SN2019oys was first detected in the optical and was classified as a Type Ib SN. Then, about $\sim 100$ days after discovery, it showed an optical rebrightening and a spectral transition to a spectrum dominated by strong narrow emission lines, which suggests strong interaction with

Roman Parzer, Peter Filzmoser, Laura Vana-Gür

We examine the linear regression problem in a challenging high-dimensional setting with correlated predictors where the vector of coefficients can vary from sparse to dense. In this setting, we propose a combination of probabilistic variable screening with random projection tools as a viable approach. More specifically, we introduce a new data-driven random

Miha Nemevšek, Yue Zhang

Temporary matter domination and late entropy dilution, injected by a "long-lived" particle in the early universe, serves as a standard mechanism for yielding the correct dark matter relic density. We recently pointed out the cosmological significance of diluting particle's partial decay into dark matter. When repopulated in such a way, dark matter carries hi

Yumou Wei, Ryan F. Forelli, Chris Hansen, Jeffrey P. Levesque

Active feedback control in magnetic confinement fusion devices is desirable to mitigate plasma instabilities and enable robust operation. Optical high-speed cameras provide a powerful, non-invasive diagnostic and can be suitable for these applications. In this study, we process fast camera data, at rates exceeding 100kfps, on $\textit{in situ}$ Field Program

Andrea L. Guerrieri, Aditya Hebbar, Balt C. van Rees

We set up a numerical S-matrix bootstrap problem to rigorously constrain bound state couplings given by the residues of poles in elastic amplitudes. We extract upper bounds on these couplings that follow purely from unitarity, crossing symmetry, and the Roy equations within their proven domain of validity. First we consider amplitudes with a single spin 0 or

Lucian Beznea, Alexandra Teodor

We give a probabilistic representation of the solution to a semilinear elliptic Dirichlet problem with general (discontinuous) boundary data. The boundary behaviour of the solution is in the sense of the controlled convergence initiated by A. Cornea. Uniqueness results for the solution are also provided.

Tobías I. Liaudat, Matthijs Mars, Matthew A. Price, Marcelo Pereyra

Next-generation radio interferometers like the Square Kilometer Array have the potential to unlock scientific discoveries thanks to their unprecedented angular resolution and sensitivity. One key to unlocking their potential resides in handling the deluge and complexity of incoming data. This challenge requires building radio interferometric imaging methods

P. Tzanavaris, T. Yaqoob, S. LaMassa

Narrow Fe K$\alpha$ fluorescent emission lines arising at $\sim$kpc-scale separations from the nucleus have only been detected in a few AGN. The detections require that the extended line emission be spatially resolved and sufficiently bright. Compared to narrow Fe K$\alpha$ lines arising closer to the nucleus, they have much lower fluxes but show substantial

Joschka Birk, Erik Buhmann, Cedric Ewen, Gregor Kasieczka

We introduce the first generative model trained on the JetClass dataset. Our model generates jets at the constituent level, and it is a permutation-equivariant continuous normalizing flow (CNF) trained with the flow matching technique. It is conditioned on the jet type, so that a single model can be used to generate the ten different jet types of JetClass. F

D. B. Jess, S. D. T. Grant, W. Bate, J. Liu

The solar physics community is entering a golden era that is ripe with next-generation ground- and space-based facilities. With ever-increasing resolving power stemming from the newest observational telescopes, it becomes more challenging to obtain (near-)simultaneous measurements at high spatial, temporal and spectral resolutions, while operating at the dif

Vladimir Strokov, Emanuele Berti

Among the binary sources of interest for LISA some are quasimonochromatic, in the sense that the change in the gravitational wave frequency $\Delta f\lesssim 1\;\mbox{yr}^{-1}$ during the observation time. For these sources, we revisit the stationary phase approximation (SPA) commonly used in Fisher matrix calculations in the frequency domain and show how it

Tom Rudelius

Castellano, Ruiz, and Valenzuela recently observed a remarkable "pattern" in infinite-distance limits of moduli spaces in quantum gravity, which relates the field space variation of the mass of the lightest tower of particles to the field space variation of the species scale. In this work, we show how a version of this pattern can be proven to hold for BPS p

Eric M. Metodiev, Jesse Thaler, Raymond Wynne

To maximize the discovery potential of high-energy colliders, experimental searches should be sensitive to unforeseen new physics scenarios. This goal has motivated the use of machine learning for unsupervised anomaly detection. In this paper, we introduce a new anomaly detection strategy called FORCE: factorized observables for regressing conditional expect

Mesut Çalışkan, Neha Anil Kumar, Selim C. Hotinli, Marc Kamionkowski

The intergalactic helium became fully ionized by the end of cosmic noon ($z\sim2$). Similarly to the reionization of hydrogen, helium reionization is expected to be patchy, driven by luminous quasars that ionize the intergalactic gas in their surrounding environment. Probing the morphology of ionized electrons during this epoch can provide crucial informatio

Gege Gao, Weiyang Liu, Anpei Chen, Andreas Geiger

As pretrained text-to-image diffusion models become increasingly powerful, recent efforts have been made to distill knowledge from these text-to-image pretrained models for optimizing a text-guided 3D model. Most of the existing methods generate a holistic 3D model from a plain text input. This can be problematic when the text describes a complex scene with

Amedeo Romagnolo, Alex C. Gormaz-Matamala, Krzysztof Belczynski

In high metallicity environments the mass that black holes (BHs) can reach just after core-collapse widely depends on how much mass their progenitor stars lose via winds. On one hand new theoretical and observational insights suggest that early-stage winds should be weaker than what many canonical models prescribe. On the other hand the proximity to the Eddi

Dominick Reilly, Srijan Das

Video transformers have become the de facto standard for human action recognition, yet their exclusive reliance on the RGB modality still limits their adoption in certain domains. One such domain is Activities of Daily Living (ADL), where RGB alone is not sufficient to distinguish between visually similar actions, or actions observed from multiple viewpoints

Lihao Liu, Yanqi Cheng, Zhongying Deng, Shujun Wang

Multi-object tracking in traffic videos is a crucial research area, offering immense potential for enhancing traffic monitoring accuracy and promoting road safety measures through the utilisation of advanced machine learning algorithms. However, existing datasets for multi-object tracking in traffic videos often feature limited instances or focus on single c

Zeyuan Yin, Zhiqiang Shen

Dataset distillation or condensation aims to generate a smaller but representative subset from a large dataset, which allows a model to be trained more efficiently, meanwhile evaluating on the original testing data distribution to achieve decent performance. Previous decoupled methods like SRe$^2$L simply use a unified gradient update scheme for synthesizing

Zhen Xing, Qi Dai, Zihao Zhang, Hui Zhang

Diffusion models have achieved significant success in image and video generation. This motivates a growing interest in video editing tasks, where videos are edited according to provided text descriptions. However, most existing approaches only focus on video editing for short clips and rely on time-consuming tuning or inference. We are the first to propose V

Yao Feng, Jing Lin, Sai Kumar Dwivedi, Yu Sun

We introduce ChatPose, a framework employing Large Language Models (LLMs) to understand and reason about 3D human poses from images or textual descriptions. Our work is motivated by the human ability to intuitively understand postures from a single image or a brief description, a process that intertwines image interpretation, world knowledge, and an understa

Jackson Van Dyke

This paper shows that quantization of $\pi$-finite spaces, as a functor out of a higher category of spans, is equivariant in two ways: Symmetries of a given polarization/Lagrangian always induce coherent symmetries of the quantization. On the other hand, symmetries of the entire phase space a priori only induce projective symmetries, with an invertible once-

Rongyao Fang, Shilin Yan, Zhaoyang Huang, Jingqiu Zhou

Empowering models to dynamically accomplish tasks specified through natural language instructions represents a promising path toward more capable and general artificial intelligence. In this work, we introduce InstructSeq, an instruction-conditioned multi-modal modeling framework that unifies diverse vision tasks through flexible natural language control and

Or Greenberg, Eran Kishon, Dani Lischinski

Image-to-image translation (I2IT) refers to the process of transforming images from a source domain to a target domain while maintaining a fundamental connection in terms of image content. In the past few years, remarkable advancements in I2IT were achieved by Generative Adversarial Networks (GANs), which nevertheless struggle with translations requiring hig

Wenming Weng, Ruoyu Feng, Yanhui Wang, Qi Dai

We present ART$\boldsymbol{\cdot}$V, an efficient framework for auto-regressive video generation with diffusion models. Unlike existing methods that generate entire videos in one-shot, ART$\boldsymbol{\cdot}$V generates a single frame at a time, conditioned on the previous ones. The framework offers three distinct advantages. First, it only learns simple con

Andrew J. Sargent, Travis C. Fischer, Megan C. Johnson, Alexander J. van der Horst

We present the first results from Karl G. Jansky Very Large Array (VLA) observations as a part of the Fundamental Reference Active Galactic Nucleus (AGN) Monitoring Experiment (FRAMEx), a program to understand the relationship between AGN accretion physics and wavelength-dependent position as a function of time. With this VLA survey, we investigate the radio

Matthew Gwilliam, Michael Cogswell, Meng Ye, Karan Sikka

Existing long video retrieval systems are trained and tested in the paragraph-to-video retrieval regime, where every long video is described by a single long paragraph. This neglects the richness and variety of possible valid descriptions of a video, which could range anywhere from moment-by-moment detail to a single phrase summary. To provide a more thoroug

Hsin-Ying Lee, Hung-Yu Tseng, Hsin-Ying Lee, Ming-Hsuan Yang

Contents generated by recent advanced Text-to-Image (T2I) diffusion models are sometimes too imaginative for existing off-the-shelf dense predictors to estimate due to the immitigable domain gap. We introduce DMP, a pipeline utilizing pre-trained T2I models as a prior for dense prediction tasks. To address the misalignment between deterministic prediction ta

Hidekazu Tanaka, Shuji Sasagawa

We extend the Schwinger-Dyson equation (SDE) on the complex plane, which was treated in our previous research, to finite temperature. As a simple example, we solve the SDE for a model with four-fermion interactions in the (1+1) space-time dimensions at strong coupling region. We investigate the properties of the effective mass and energy for the fermions, es

Shuyuan Tu, Qi Dai, Zhi-Qi Cheng, Han Hu

Existing diffusion-based video editing models have made gorgeous advances for editing attributes of a source video over time but struggle to manipulate the motion information while preserving the original protagonist's appearance and background. To address this, we propose MotionEditor, a diffusion model for video motion editing. MotionEditor incorporates a

Ziyun Wang, Jinyuan Guo, Kostas Daniilidis

Event cameras are a novel type of biologically inspired vision sensor known for their high temporal resolution, high dynamic range, and low power consumption. Because of these properties, they are well-suited for processing fast motions that require rapid reactions. Although event cameras have recently shown competitive performance in unsupervised optical fl

Yanhui Wang, Jianmin Bao, Wenming Weng, Ruoyu Feng

We present MicroCinema, a straightforward yet effective framework for high-quality and coherent text-to-video generation. Unlike existing approaches that align text prompts with video directly, MicroCinema introduces a Divide-and-Conquer strategy which divides the text-to-video into a two-stage process: text-to-image generation and image\&text-to-video gener

Ziyun Wang, Friedhelm Hamann, Kenneth Chaney, Wen Jiang

We present ContinuityCam, a novel approach to generate a continuous video from a single static RGB image and an event camera stream. Conventional cameras struggle with high-speed motion capture due to bandwidth and dynamic range limitations. Event cameras are ideal sensors to solve this problem because they encode compressed change information at high tempor

Tianwei Yin, Michaël Gharbi, Richard Zhang, Eli Shechtman

Diffusion models generate high-quality images but require dozens of forward passes. We introduce Distribution Matching Distillation (DMD), a procedure to transform a diffusion model into a one-step image generator with minimal impact on image quality. We enforce the one-step image generator match the diffusion model at distribution level, by minimizing an ap

Agelos Kratimenos, Jiahui Lei, Kostas Daniilidis

Accurately and efficiently modeling dynamic scenes and motions is considered so challenging a task due to temporal dynamics and motion complexity. To address these challenges, we propose DynMF, a compact and efficient representation that decomposes a dynamic scene into a few neural trajectories. We argue that the per-point motions of a dynamic scene can be d

Wilson Yan, Andrew Brown, Pieter Abbeel, Rohit Girdhar

We introduce MoCA, a Motion-Conditioned Image Animation approach for video editing. It leverages a simple decomposition of the video editing problem into image editing followed by motion-conditioned image animation. Furthermore, given the lack of robust evaluation datasets for video editing, we introduce a new benchmark that measures edit capability across a

Kaiwen Hou

This paper presents a groundbreaking approach to causal inference by integrating continuous normalizing flows (CNFs) with parametric submodels, enhancing their geometric sensitivity and improving upon traditional Targeted Maximum Likelihood Estimation (TMLE). Our method employs CNFs to refine TMLE, optimizing the Cram\'er-Rao bound and transitioning from a p

Dongho Lee, Jongseo Lee, Jinwoo Choi

Recognizing human actions in videos requires spatial and temporal understanding. Most existing action recognition models lack a balanced spatio-temporal understanding of videos. In this work, we propose a novel two-stream architecture, called Cross-Attention in Space and Time (CAST), that achieves a balanced spatio-temporal understanding of videos using only

Kyungho Bae, Geo Ahn, Youngrae Kim, Jinwoo Choi

Video recognition models often learn scene-biased action representation due to the spurious correlation between actions and scenes in the training data. Such models show poor performance when the test data consists of videos with unseen action-scene combinations. Although scene-debiased action recognition models might address the issue, they often overlook v

Alexander Downey, Evren Tuna, Alkan Soysal

We have developed a new framework using time-series analysis for dynamically assigning mobile network traffic prediction models in previously unseen wireless environments. Our framework selectively employs learned behaviors, outperforming any single model with over a 50% improvement relative to current studies. More importantly, it surpasses traditional appr

Zhiqiu Xu, Yanjie Chen, Kirill Vishniakov, Yida Yin

Weight initialization plays an important role in neural network training. Widely used initialization methods are proposed and evaluated for networks that are trained from scratch. However, the growing number of pretrained models now offers new opportunities for tackling this classical problem of weight initialization. In this work, we introduce weight select

Moayed Haji-Ali, Guha Balakrishnan, Vicente Ordonez

Diffusion models have revolutionized image generation in recent years, yet they are still limited to a few sizes and aspect ratios. We propose ElasticDiffusion, a novel training-free decoding method that enables pretrained text-to-image diffusion models to generate images with various sizes. ElasticDiffusion attempts to decouple the generation trajectory of

J. J. Blanco-Pillado, J. R. Espinosa, J. Huertas, K. Sousa

The catastrophic decay of a spacetime with compact dimensions, via bubbles of nothing (BoNs), is probably a generic phenomenon. BoNs admit a 4-dimensional description as singular Coleman-de Luccia bounces of the size modulus field, stabilized by some potential $V(\phi)$. We apply the tunneling potential approach to this 4d description to provide a very simpl

Sujata Sinha, Alkan Soysal

We introduce a Channel Distribution Information (CDI)-aware Generative Adversarial Network (GAN), designed to address the unique challenges of adversarial attacks in wireless communication systems. The generator in this CDI-aware GAN maps random input noise to the feature space, generating perturbations intended to deceive a target modulation classifier. Its

Nanxi Yao, Agnieszka Sorensen, Veronica Dexheimer, Jacquelyn Noronha-Hostler

From the observation of both heavy neutron stars and light ones with small radii, one anticipates a steep rise in the speed of sound of nuclear matter as a function of baryon density up to values close to the causal limit. A question follows whether such behavior of the speed of sound in neutron-rich matter is compatible with the equation of state extracted

Rouven Frassek, Alexander Tsymbaliuk

We study the RTT orthosymplectic super Yangians and present their Drinfeld realizations for any parity sequence, generalizing the results for non-super types BCD, a standard parity sequence, and super A-type.

Kaifeng Lyu, Jikai Jin, Zhiyuan Li, Simon S. Du

Recent work by Power et al. (2022) highlighted a surprising "grokking" phenomenon in learning arithmetic tasks: a neural net first "memorizes" the training set, resulting in perfect training accuracy but near-random test accuracy, and after training for sufficiently longer, it suddenly transitions to perfect test accuracy. This paper studies the grokking phe

Zhaoning Wang, Ming Li, Chen Chen

With the recent development of generative models, Text-to-3D generations have also seen significant growth, opening a door for creating video-game 3D assets from a more general public. Nonetheless, people without any professional 3D editing experience would find it hard to achieve precise control over the 3D generation, especially if there are multiple objec

Mohammed Alzergani

We introduce an infinite family of approximations for a Dirichlet $L$-function $L(s, \chi)$ arising from truncated Euler products. These approximations are entire functions and satisfy the same functional equation as $L(s, \chi)$. We provide numerical evidence of the accuracy of estimating values of $L(s, \chi)$ in the critical strip using these approximatio

Amber Yijia Zheng, Raymond A. Yeh

Advancements in open-sourced text-to-image models and fine-tuning methods have led to the increasing risk of malicious adaptation, i.e., fine-tuning to generate harmful/unauthorized content. Recent works, e.g., Glaze or MIST, have developed data-poisoning techniques which protect the data against adaptation methods. In this work, we consider an alternative p

Yudong Wang, Jichang Guo, Wanru He, Huan Gao

Underwater object detection is a crucial and challenging problem in marine engineering and aquatic robot. The difficulty is partly because of the degradation of underwater images caused by light selective absorption and scattering. Intuitively, enhancing underwater images can benefit high-level applications like underwater object detection. However, it is st

Francesco Mori, L. Mahadevan

When navigating complex environments, animals often combine multiple strategies to mitigate the effects of external disturbances. These modalities often correspond to different sources of information, leading to speed-accuracy trade-offs. Inspired by the intermittent reorientation strategy seen in the behavior of the dung beetle, we consider the problem of t

Raphael Tang, Xinyu Zhang, Jimmy Lin, Ferhan Ture

Do large language models (LLMs) exhibit sociodemographic biases, even when they decline to respond? To bypass their refusal to "speak," we study this research question by probing contextualized embeddings and exploring whether this bias is encoded in its latent representations. We propose a logistic Bradley-Terry probe which predicts word pair preferences of

Bardienus P. Duisterhof, Zhao Mandi, Yunchao Yao, Jia-Wei Liu

Teaching robots to fold, drape, or reposition deformable objects such as cloth will unlock a variety of automation applications. While remarkable progress has been made for rigid object manipulation, manipulating deformable objects poses unique challenges, including frequent occlusions, infinite-dimensional state spaces and complex dynamics. Just as object p

Philipp Böer, Patrick Hager, Matthias Neubert, Michel Stillger

The resummation of the ''Glauber series'' in non-global LHC observables is extended to processes with gluons in the initial state. This series simultaneously incorporates large double-logarithmic corrections, the so-called ''super-leading logarithms'', together with higher-order exchanges of pairs of Glauber gluons associated with the large numerical factor

Chicago Park, Shirin Shoushtari, Weijie Gan, Ulugbek S. Kamilov

Plug-and-Play Alternating Direction Method of Multipliers (PnP-ADMM) is a widely-used algorithm for solving inverse problems by integrating physical measurement models and convolutional neural network (CNN) priors. PnP-ADMM has been theoretically proven to converge for convex data-fidelity terms and nonexpansive CNNs. It has however been observed that PnP-AD

Evin Pınar Örnek, Yann Labbé, Bugra Tekin, Lingni Ma

We propose FoundPose, a model-based method for 6D pose estimation of unseen objects from a single RGB image. The method can quickly onboard new objects using their 3D models without requiring any object- or task-specific training. In contrast, existing methods typically pre-train on large-scale, task-specific datasets in order to generalize to new objects an

Scott Balchin, Tobias Barthel, J. P. C. Greenlees

We study the tensor-triangular geometry of the category of rational $G$-spectra for a compact Lie group $G$. In particular, we prove that this category can be naturally decomposed into local factors supported on individual subgroups, each of which admits an algebraic model. This is an important step and strong evidence towards the third author's conjecture t

Jake M. Hofman, Angelos Chatzimparmpas, Amit Sharma, Duncan J. Watts

Amid rising concerns of reproducibility and generalizability in predictive modeling, we explore the possibility and potential benefits of introducing pre-registration to the field. Despite notable advancements in predictive modeling, spanning core machine learning tasks to various scientific applications, challenges such as overlooked contextual factors, dat

Akshay Punjabi, Pablo Izquierdo Ayala

Accurate precipitation forecasting is indispensable for informed decision-making across various industries. However, the computational demands of current models raise environmental concerns. We address the critical need for accurate precipitation forecasting while considering the environmental impact of computational resources and propose a minimalist U-Net

Qi Cao, Takeshi Kojima, Yutaka Matsuo, Yusuke Iwasawa

While Large Language Models (LLMs) have achieved remarkable performance in many tasks, much about their inner workings remains unclear. In this study, we present novel experimental insights into the resilience of LLMs, particularly GPT-4, when subjected to extensive character-level permutations. To investigate this, we first propose the Scrambled Bench, a su

Andrew R. Frey, Michael P. Grehan, Prakriti Singh

We study the complexity of the gravity dual to the confining $SU(N)\times SU(N+M)$ Klebanov-Strassler gauge theory, which is an important test case for holographic complexity in higher-dimensional and nonconformal gauge/gravity dualities. We emphasize the dependence of the complexity on parameters of the gauge theory, finding a common behavior with confineme

Samuel Stevens, Jiaman Wu, Matthew J Thompson, Elizabeth G Campolongo

Images of the natural world, collected by a variety of cameras, from drones to individual phones, are increasingly abundant sources of biological information. There is an explosion of computational methods and tools, particularly computer vision, for extracting biologically relevant information from images for science and conservation. Yet most of these are

N. Aimar, T. Paumard, F. H. Vincent, E. Gourgoulhon

Polarized general-relativistic radiative transfer in the vicinity of black holes and other compact objects has become a crucial tool for probing the properties of relativistic astrophysics plasmas. Instruments like GRAVITY, the Event Horizon telescope, ALMA, or IXPE make it very timely to develop such numerical frameworks. In this article, we present the pol

Ayush Baid, John Lambert, Travis Driver, Akshay Krishnan

While initial approaches to Structure-from-Motion (SfM) revolved around both global and incremental methods, most recent applications rely on incremental systems to estimate camera poses due to their superior robustness. Though there has been tremendous progress in SfM `front-ends' powered by deep models learned from data, the state-of-the-art (incremental)

Dibya Kirti Mishra, Srinjana Routh, Bibhuti Kumar Jha, Theodosios Chatzistergos

Chromospheric differential rotation is a key component in comprehending the atmospheric coupling between the chromosphere and the photosphere at different phases of the solar cycle. In this study, we therefore utilize the newly calibrated multidecadal Ca II K spectroheliograms (1907-2007) from the Kodaikanal Solar Observatory (KoSO) to investigate the differ

Artemis Panagopoulou, Le Xue, Ning Yu, Junnan Li

Recent research has achieved significant advancements in visual reasoning tasks through learning image-to-language projections and leveraging the impressive reasoning abilities of Large Language Models (LLMs). This paper introduces an efficient and effective framework that integrates multiple modalities (images, 3D, audio and video) to a frozen LLM and demon

Jishu Das

Let $S_k(N)$ denote the space of cusp forms of even integer weight $k$ and level $N$. We prove an asymptotic for the Petersson trace formula for $S_k(N)$ under an appropriate condition. Using the non-vanishing of a Kloosterman sum involved in the asymptotic, we give a lower bound for discrepancy in the Sato-Tate distribution for levels not divisible by $8$.

Hanmeng Zhan

We study whether the probability distribution of a discrete quantum walk can get arbitrarily close to uniform, given that the walk starts with a uniform superposition of the outgoing arcs of some vertex. We establish a characterization of this phenomenon on regular non-bipartite graphs in terms of their adjacency eigenvalues and eigenprojections. Using theor

Letícia Mattos, Domenico Mergoni Cecchelli, Olaf Parczyk

Schur's theorem states that in any $k$-colouring of the set of integers $[n]$ there is a monochromatic solution to $a+b=c$, provided $n$ is sufficiently large. Abbott and Wang studied the size of the largest subset of $[n]$ such that there is a $k$-colouring avoiding a monochromatic $a+b=c$. In other directions, the minimum number of $a+b=c$ in $k$-colouring

Christopher Alexander, Mahir Hadžić, Matthew Schrecker

We show the existence of a new class of initially smooth spherically symmetric self-similar solutions to the non-isentropic Euler-Poisson system. These solutions exhibit supersonic gravitational implosion in the sense that the density blows-up in finite time while the fluid velocity remains supersonic. In particular, they occupy a portion of the phase space

Yang Bai, Carlos Henrique de Lima

Axion-like particles are predicted in many physics scenarios beyond the Standard Model (SM). Their interactions with SM particles may arise from the triangle anomaly of the associated global symmetry, along with other SM global and gauge symmetries, including anomalies with the global baryon number and electromagnetic gauge symmetries. We initiate the phenom

André Baião Raposo, Maxwell T. Hansen

The two-particle finite-volume scattering formalism derived by L\"uscher and generalized in many subsequent works does not hold for energies far enough below the two-particle threshold to reach the nearest left-hand cut. The breakdown of the formalism is signaled by the fact that a real scattering amplitude is predicted in a regime where it should be complex

Ahmed S. Alahmed, Lang Tong

We analyze the overall benefits of an energy community cooperative game under which distributed energy resources (DER) are shared behind a regulated distribution utility meter under a general net energy metering (NEM) tariff. Two community DER scheduling algorithms are examined. The first is a community with centrally controlled DER, whereas the second is de

Ege Orkun Gamgam, Nail Akar, Sennur Ulukus

We study the scheduling problem for a multi-source single-server generate-at-will (GAW) status update system with sources having heterogeneous service times and weights, with the goal of minimizing the weighted sum age of information (AoI). In particular, we study \emph{age-agnostic} schedulers which rely only on the first two moments of the source service t

Manuel D. Contreras, Carlos Gómez-Cabello, Luis Rodríguez-Piazza

The algebra of Dirichlet series $\mathcal{A}(\mathcal{C}_{+})$ consists on those Dirichlet series convergent in the right half-plane $\mathcal{C}_{+}$ and which are also uniformly continuous there. This algebra was recently introduced by Aron, Bayart, Gauthier, Maestre, and Nestoridis. We describe the symbols $\Phi:\mathcal{C}_{+}\to\mathcal{C}_{+}$ giving r