February 2025 arXiv papers

Lucas Seiffert, Felipe Pereira

We consider the theoretical analysis of Multiscale Sampling Methods, which are a new class of gradient-free Markov chain Monte Carlo (MCMC) methods for high dimensional inverse differential equation problems. A detailed presentation of those methods is given, including a review of each MCMC technique that they employ. Then, we propose a two-part framework to

Lukas William Mayer, Sheer Karny, Jackie Ayoub, Miao Song

Despite the growing interest in collaborative AI, designing systems that seamlessly integrate human input remains a major challenge. In this study, we developed a task to systematically examine human preferences for collaborative agents. We created and evaluated five collaborative AI agents with strategies that differ in the manner and degree they adapt to h

Dac-Nhan-Tam Nguyen

Let $p$ be an odd prime. Consider normalized newforms $f_1,f_2$ that both satisfy the Heegner hypothesis for an imaginary quadratic field $K$ and suppose that they induce isomorphic residual Galois representations. In the work of Greenberg-Vatsal and Emerton-Pollack-Weston, the authors compare the cyclotomic Iwasawa $μ$ and $λ$-invariants of $f_1$ and $f_2$.

Michał Wichrowski

We present a matrix-free approach for implementing ghost penalty stabilization in Cut Finite Element Methods (CutFEM). While matrix-free methods for CutFEM have been developed, the efficient evaluation of high-order, face-based ghost penalties remains a significant challenge, which this work addresses. By exploiting the tensor-product structure of the ghost

K. K. Ernazarov, V. D. Ivashchuk

We consider the $4D$ gravitational model with a scalar field $φ$, Einstein and Gauss-Bonnet terms. The action of the model contains a potential term $U(φ)$, Gauss-Bonnet coupling function $f(φ)$ and a parameter $\varepsilon = \pm1 $, where $\varepsilon = 1$ corresponds to ordinary scalar field and $\varepsilon = -1 $ - to phantom one. Inspired by the recent

Martino Grasselli, Gilles Pagès

We investigate a class of non-Markovian processes that hold particular relevance in the realm of mathematical finance. This family encompasses path-dependent volatility models, including those pioneered by [Platen and Rendek, 2018] and, more recently, by [Guyon and Lekeufack, 2023]. Our study unfolds in two principal phases. In the first phase, we introduce

Marius F. R. Juston, Ramavarapu S. Sreenivas, William R. Norris, Dustin Nottage

This paper discusses the weight parametrization of two standard 1-Lipschitz network architectures, the Almost-Orthogonal-Layers (AOL) and the SDP-based Lipschitz Layers (SLL). It examines their impact on initialization for deep 1-Lipschitz feedforward networks, and discusses underlying issues surrounding this initialization. These networks are mainly used in

Michal Spiegel, Michal Štefánik, Marek Kadlčík, Josef Kuchař

Can transformers learn to perform algorithmic tasks reliably across previously unseen input/output domains? While pre-trained language models show solid accuracy on benchmarks incorporating algorithmic reasoning, assessing the reliability of these results necessitates an ability to distinguish genuine algorithmic understanding from memorization. In this pape

Xiang-Yu Huang, Simon Birrer, Michele Cappellari, Tommaso Treu

Constraining the mass-sheet degeneracy (MSD) is crucial for improving the precision and accuracy of time-delay cosmography. Joint analyses of lensing and stellar kinematics are widely adopted to break the MSD. A 3D mass and stellar tracer population is required to accurately interpret the kinematics data. Our forward-modeling procedure aims at evaluating the

Lukasz Sztukiewicz, Ignacy Stępka, Michał Wiliński, Jerzy Stefanowski

The widespread adoption of machine learning systems has raised critical concerns about fairness and bias, making mitigating harmful biases essential for AI development. In this paper, we investigate the relationship between debiasing and removing artifacts in neural networks for computer vision tasks. First, we introduce a set of novel XAI-based metrics that

Wenqi Huang, Nan Wang, Congyu Liao, Yimeng Lin

Echo Planar Imaging (EPI) is widely used for its rapid acquisition but suffers from severe geometric distortions due to B0 inhomogeneities, particularly along the phase encoding direction. Existing methods follow a two-step process: reconstructing blip-up/down EPI images, then estimating B0, which can introduce error accumulation and reduce correction accura

Sharan Vaswani, Reza Babanezhad

Armijo line-search (Armijo-LS) is a standard method to set the step-size for gradient descent (GD). For smooth functions, Armijo-LS alleviates the need to know the global smoothness constant L and adapts to the ``local'' smoothness, enabling GD to converge faster. Existing theoretical analyses show that GD with Armijo-LS (GD-LS) can result in constan

Melih İşeri, Erhan Bayraktar

This work introduces a unified framework for analyzing games in greater depth. In the existing literature, players' strategies are typically assigned scalar values, and equilibrium concepts are used to identify compatible choices. However, this approach neglects the internal structure of players, thereby failing to accurately model observed behaviors. To

Yufei Guo, Xiaode Liu, Yuanpei Chen, Weihang Peng

Transformers have demonstrated outstanding performance across a wide range of tasks, owing to their self-attention mechanism, but they are highly energy-consuming. Spiking Neural Networks have emerged as a promising energy-efficient alternative to traditional Artificial Neural Networks, leveraging event-driven computation and binary spikes for information tr

Pengcheng Jiang, Jiacheng Lin, Lang Cao, Runchu Tian

Information retrieval systems are crucial for enabling effective access to large document collections. Recent approaches have leveraged Large Language Models (LLMs) to enhance retrieval performance through query augmentation, but often rely on expensive supervised learning or distillation techniques that require significant computational resources and hand-l

Le Minh Anh Nguyen, Brant Bowers, Sara Mouradian

To increase the power of a trapped ion quantum information processor, the qubit number, gate speed, and gate fidelity must all increase. All three of these parameters are influenced by the trapping field which in turn depends on the electrode geometry. Here we consider how the electrode geometry affects the radial trapping parameters: trap height, harmonicit

Reza Saadati, Fayçal Hammad

We recently studied spin precession in various stationary and axisymmetric spacetimes and applied it to the case of neutrinos propagating in those spacetimes. In this paper, the study of spin precession is extended to the case of a spinning particle propagating within the spacetime of a time-dependent gravitational plane wave. First, the angular velocity of

Abhishek Setty, Lukas Morand, Poojitha Ramachandra, Claas Bierwisch

This study explores the applicability of a graph-based interaction-aware trajectory prediction model, originally developed for the transportation domain, to forecast particle trajectories in three-dimensional discrete element simulations. The model and our enhancements are validated at two typical particle simulation use cases: (i) particle flow in a represe

Alex Alberts, Ilias Bilionis

Many inverse problems require reconstructing physical fields from limited and noisy data while incorporating known governing equations. A growing body of work within probabilistic numerics formalizes such tasks via Bayesian inference in function spaces by assigning a physically meaningful prior to the latent field. In this work, we demonstrate that Brownian

Jesús Jerónimo-Castro, Luis Montejano, Efrén Morales-Amaya

We prove several localized characterizations of quadrics and, between them, the localized Blaschke's Theorem and use this result to give some characterizations of the ellipsoid related to the flatness of its grazes.

Jiawei Zhang, Xuan Yang, Taiqi Wang, Yu Yao

Traditional autonomous driving systems often struggle to connect high-level reasoning with low-level control, leading to suboptimal and sometimes unsafe behaviors. Recent advances in multimodal large language models (MLLMs), which process both visual and textual data, offer an opportunity to unify perception and reasoning. However, effectively embedding prec

S M Taslim Uddin Raju

As the population of older adults increases, so will the need for both human and robot care providers. While traditional practices involve hiring human caregivers to serve meals and attend to basic needs, older adults often require continuous companionship and health monitoring. However, hiring human caregivers for this job costs a lot of money. However, usi

Naveen Mysore

Reinforcement learning (RL) methods frequently assume that each new observation completely reflects the environment's state, thereby guaranteeing Markovian (one-step) transitions. In practice, partial observability or sensor/actuator noise often invalidates this assumption. This paper proposes a systematic methodology for detecting such violations, combi

Mikołaj Rogóż, Zofia Dziekan, Piotr Wasylczyk

Depending on multiple parameters, soft robots can exhibit different modes of locomotion that are difficult to model numerically. As a result, improving their performance is complex, especially in small-scale systems characterized by low Reynolds numbers, when multiple aero- and hydrodynamical processes influence their movement. In this work, we optimize ligh

William Nguyen, An Phan, Konobu Kimura, Hitoshi Maeno

Large Language Models (LLMs) have demonstrated substantial potential in addressing complex reasoning tasks, yet their general-purpose nature often limits their effectiveness in specialized domains such as maritime navigation. To bridge this gap, we introduce Llamarine, the first open-source LLM designed specifically for maritime navigation. Llamarine 1.0 is

Shuang Li, Yihuai Gao, Dorsa Sadigh, Shuran Song

A unified video and action model holds significant promise for robotics, where videos provide rich scene information for action prediction, and actions provide dynamics information for video prediction. However, effectively combining video generation and action prediction remains challenging, and current video generation-based methods struggle to match the p

Jacob M. Hiesener, C. Alex Kaylor, Joshua J. Wong, Prankush Agarwal

We present a seeded topology optimization methodology for integrated photonic devices fabricated on foundry platforms that yields improved performance compared to traditional topology optimization. We employ blurring filters and a design rule check correction algorithm to more readily meet fabrication constraints, resulting in devices with fewer artifacts an

Martin Hartenberger, Huzeyfe Ayaz, Fatih Ozlugedik, Charly Caredda

In this paper, we propose a methodology for extracting molecular tumor biomarkers from hyperspectral imaging (HSI), an emerging technology for intraoperative tissue assessment. To achieve this, we employ spectral unmixing, allowing to decompose the spectral signals recorded by the HSI camera into their constituent molecular components. Traditional unmixing a

Amar Kumar, Anita Kriz, Mohammad Havaei, Tal Arbel

Developing reliable and generalizable deep learning systems for medical imaging faces significant obstacles due to spurious correlations, data imbalances, and limited text annotations in datasets. Addressing these challenges requires architectures that are robust to the unique complexities posed by medical imaging data. Rapid advancements in vision-language

Jiawei Zhang, Shuang Yang, Bo Li

Large Language Model (LLM) agents equipped with external tools have become increasingly powerful for complex tasks such as web shopping, automated email replies, and financial trading. However, these advancements amplify the risks of adversarial attacks, especially when agents can access sensitive external functionalities. Nevertheless, manipulating LLM agen

Paula Ebner, Jessica Szczuka

LLM-based chatbots are now being specifically designed to facilitate social companionship, even romantic relationships, incorporating features that parallel human relationship dynamics. This has led a subset of users to form romantic relationships with chatbots. Understanding which interpersonal characteristics drive individuals to form intense, emotional bo

Md Nur-A-Adam Dony

Reinforcement learning (RL) has emerged as a potent paradigm for autonomous decision-making in complex environments. However, the integration of event-driven decision processes within RL remains a challenge. This paper presents a novel architecture that combines a Discrete Event Supervisory (DES) model with a standard RL framework to create a hybrid decision

Xinhang Ma, Junlin Wu, Hussein Sibai, Yiannis Kantaros

Ensuring safety in autonomous systems with vision-based control remains a critical challenge due to the high dimensionality of image inputs and the fact that the relationship between true system state and its visual manifestation is unknown. Existing methods for learning-based control in such settings typically lack formal safety guarantees. To address this

Matteo Boselli, Joel Grebel, Ambroise Peugeot, Rémy Dassonneville

Amplifying microwave signals with a noise close to the minimum imposed by quantum mechanics is now routinely performed with superconducting quantum devices. In particular, Josephson-based Traveling Wave Parametric Amplifiers (JTWPA) have shown record bandwidth with added noise close to the quantum limit. In this work, we report the appearance of echo signals

Hanjiang Hu, Alexander Robey, Changliu Liu

Large language models (LLMs) are shown to be vulnerable to jailbreaking attacks where adversarial prompts are designed to elicit harmful responses. While existing defenses effectively mitigate single-turn attacks by detecting and filtering unsafe inputs, they fail against multi-turn jailbreaks that exploit contextual drift over multiple interactions, gradual

Jorge Fariña-Asategui, Santiago Radi

We prove that super strongly fractal groups acting on regular rooted trees have null fixed-point proportion. In particular, we show that the fixed-point proportion of an infinite family of iterated monodromy groups of exceptional complex polynomials have the same property. The proof uses the approach of Rafe Jones in [15] based on martingales and a recent re

Michael Park, Erin Leahey, Russell J. Funk

We respond to Holst et al.'s critique that the decline in scientific disruptiveness documented in Park et al. (Nature, 2023) is an artifact of including works with zero backward citations. Using their advocated dataset, metric, and exclusion criteria, we find declines equivalent to major benchmark transformations in science. Their own regression model--d

Jeffrey D. Adler, Joshua M. Lansky, Loren Spice

Let $k$ be a field, and suppose that $Γ$ is a smooth $k$-group that acts on a connected, reductive $k$-group $\widetilde G$. Let $G$ denote the maximal smooth, connected subgroup of the group of $Γ$-fixed points in $\widetilde G$. Under fairly general conditions, we show that $G$ is a reductive $k$-group, and that the image of the functorial embedding $\math

Zhiqiu Xia, Jinxuan Xu, Yuqian Zhang, Hang Liu

Large language models (LLMs) have demonstrated remarkable capabilities across various tasks. However, these models could offer biased, hallucinated, or non-factual responses camouflaged by their fluency and realistic appearance. Uncertainty estimation is the key method to address this challenge. While research efforts in uncertainty estimation are ramping up

Roi Bar-Zur, Ittay Eyal

Many blockchain-based decentralized services require their validators (operators) to deposit stake (collateral), which is forfeited (slashed) if they misbehave. Restaking networks let validators secure multiple services by reusing stake. These networks have quickly gained traction, leveraging over \$20 billion in stake. However, restaking introduces a new at

Dip Joti Paul, Tony X. Zhou, Karl K. Berggren

In this work, we present a technique to determine the mid-infrared refractive indices of thin superconducting films using Fourier transform infrared spectroscopy (FTIR). In particular, we performed FTIR transmission and reflection measurements on 10-nm-thick NbN and 15-nm-thick MoSi films in the wavelength range of 2.5 to 25 $μ$m, corresponding to frequencie

Benedikt Blumenstiel, Nassim Ait Ali Braham, Conrad M Albrecht, Stefano Maurogiovanni

This work presents SSL4EO-S12 v1.1, a multimodal, multitemporal Earth Observation dataset designed for pretraining large-scale foundation models. Building on the success of SSL4EO-S12, this extension updates the previous version to fix geospatial alignment inaccuracies and the inefficent data structure. The dataset allows low-barrier, analysis-ready data loa

J. P. Cuenca, T. de Ara, A. Martinez-Garcia, F. Guzman

We present an in-depth study of electronic transport in atomic-sized gold contacts using Break-Junction (BJ) techniques under cryogenic and ambient conditions. Our experimental results, supported by classical molecular dynamics (CMD) simulations and ab initio calculations, provide compelling evidence for the formation of three-atom-thick structures in gold n

Aakash Patel, Tianqing Zhang, Camille Avestruz, Jeffrey Regier

Neural posterior estimation (NPE), a type of amortized variational inference, is a computationally efficient means of constructing probabilistic catalogs of light sources from astronomical images. To date, NPE has not been used to perform inference in models with spatially varying covariates. However, ground-based astronomical images have spatially varying s

Qianran Xia, Dragan Huterer, Nhat-Minh Nguyen

Recent observations by the James Webb Space Telescope have revealed massive galaxies at very high redshift ($z\simeq 7-15$). The question of whether the existence of such galaxies is expected in the corresponding JWST surveys has received a lot of attention, though the answer straddles areas of cosmology and complex astrophysical details of high-redshift gal

Lidia Gordo Malagón, Jesús Yepes Nicolás

In this work, the $L_p$ version (for $p> 1$) of the dimensional Brunn-Minkowski inequality for the standard Gaussian measure $γ_n(\cdot)$ on $\mathbb{R}^n$ is shown. More precisely, we prove that for any $0$-symmetric convex sets with nonempty interior, any $p>1$, and every $λ\in (0,1)$, \[ γ_n\bigl((1-λ)\cdot K+_p λ\cdot L\bigr)^{p/n} \geqslant (1-λ) γ_n(K)

Fakhraddin Alwajih, Abdellah El Mekki, Samar Mohamed Magdy, Abdelrahim A. Elmadany

As large language models (LLMs) become increasingly integrated into daily life, ensuring their cultural sensitivity and inclusivity is paramount. We introduce our dataset, a year-long community-driven project covering all 22 Arab countries. The dataset includes instructions (input, response pairs) in both Modern Standard Arabic (MSA) and dialectal Arabic (DA

Dor Polonsky

The phenomenon of stable lift oscillations occurring on an elliptic wing section utilizing circulation control at transonic speeds was evaluated using numerical simulations. As the momentum of the jet increases beyond a prescribed magnitude, periodic detachment occurs from the trailing-edge. This behavior conforms to a bi-stable state, consistent with prior

Mengzhu Katie Chen, Isabella Pedraza Pineros, Arvind Satyanarayan, Jonathan Zong

Tactile charts are essential for conveying data to blind and low vision (BLV) readers but are difficult for designers to construct. Non-expert designers face barriers to entry due to complex guidelines, while experts struggle with fragmented and time-consuming workflows that involve extensive customization. Inspired by formative interviews with expert tactil

Gebhard Boeckle, Peter Mathias Graef, Theresa Kaiser

In this article, we describe a computational study of the action of the two natural $U$-operators acting on $Γ$-invariant spaces of harmonic cocycles for $\mathrm{GL}_3$ for certain congruence subgroups $Γ$, in a positive characteristic setting. The cocycle spaces we consider are conjecturally isomorphic to spaces of Drinfeld cusp forms of rank $3$ and level

Abdessamad El-Kabid, El-Mahdi El-Mhamdi

Machine learning systems deployed in distributed or federated environments are highly susceptible to adversarial manipulations, particularly availability attacks -- rendering the trained model unavailable. Prior research in distributed ML has demonstrated such adversarial effects through the injection of gradients or data poisoning. In this work, we ask whet

Kaniba Mady Keita, Younouss Hamèye Dicko

In this manuscript, we demonstrate, using several regression techniques, that the remaining independent Hodge numbers of complete intersection Calabi-Yau four-folds and five-folds can be machine learned from $h^{1,1}$ and $h^{2,1}$. Consequently, we combine the Hodge numbers $h^{1,1}$ and $h^{2,1}$ from the complete intersection Calabi-Yau three-folds, four-

R. Loganayagam, Omkar Shetye

We extend our proposal for static patch holography made in part I (arXiv:2309.07290) to the case of electromagnetism. Using the on-shell action on de Sitter Schwinger-Keldysh (dS-SK) geometry, we derive the influence phase of an observer in dS$_{d+1}$ interacting with bulk electromagnetic fields. This influence phase, computed with appropriate boundary condi

Zhongqi Yang, Amir Rahmani

Large Language Models (LLMs) excel at general-purpose reasoning by leveraging broad commonsense knowledge, but they remain limited in tasks requiring personalized reasoning over multifactorial personal data. This limitation constrains their applicability in domains such as healthcare, where decisions must adapt to individual contexts. We introduce Personaliz

Zhixian Hu, Juan Wachs, Yu She

Tactile sensing and the manipulation of delicate objects are critical challenges in robotics. This study presents a vision-based magnetic-actuated whisker array sensor that integrates these functions. The sensor features eight whiskers arranged circularly, supported by an elastomer membrane and actuated by electromagnets and permanent magnets. A camera track

Farouk Mokhtar, Joosep Pata, Dolores Garcia, Eric Wulff

We demonstrate transfer learning capabilities in a machine-learned algorithm trained for particle-flow reconstruction in high energy particle colliders. This paper presents a cross-detector fine-tuning study, where we initially pretrain the model on a large full simulation dataset from one detector design, and subsequently fine-tune the model on a sample wit

S. Orlando, H. -T. Janka, A. Wongwathanarat, D. Dickinson

[Abridged] Recent JWST observations have revealed an intricate filamentary network of unshocked ejecta in the young supernova remnant (SNR) Cassiopeia A (Cas A), offering new insights into supernova (SN) explosions and ejecta evolution. We investigate the origin and evolution of this structure by (i) characterizing its 3D morphology and kinematics and (ii) i

Anna Beer, Lena Krieger, Pascal Weber, Martin Ritzert

Being able to evaluate the quality of a clustering result even in the absence of ground truth cluster labels is fundamental for research in data mining. However, most cluster validation indices (CVIs) do not capture noise assignments by density-based clustering methods like DBSCAN or HDBSCAN, even though the ability to correctly determine noise is crucial fo

Gabriel Rodrigues, Rayff de Souza, Jamerson Rodrigues, Jailson Alcaniz

Recent analyses have shown that a dynamic dark energy modeled by the CPL parameterization of the dark energy equation of state (EoS) can ease constraints on the total neutrino mass compared to the standard $Λ$CDM model. This helps reconcile cosmological and particle physics measurements of $\sum m_ν$. In this study, we investigate the robustness of this effe

Mehrdad Phoroutan-Mehr, Tara Fetherolf

Exoplanets, with their large volumes and low temperatures, are ideal celestial detectors for probing dark matter (DM) interactions. DM particles can lose energy through scattering with the planetary interior and become gravitationally captured if their interaction with the visible sector is sufficiently strong. In the absence of annihilation, the captured DM

Thomas Massoni

This survey explores the geometry of three-dimensional Anosov flows from the perspective of contact and symplectic geometry, following the work of Mitsumatsu, Eliashberg-Thurston, Hozoori, and the author. We also present a few original results and discuss various open questions and conjectures.

Ali Ahmad Khostovan, Jeyhan S. Kartaltepe, Mara Salvato, Olivier Ilbert

We present the COSMOS Spectroscopic Redshift Compilation encompassing ~ 20 years of spectroscopic redshifts within a 10 deg$^2$ area centered on the 2 deg$^2$ COSMOS legacy field. This compilation contains 487,666 redshifts of 266,284 unique objects from 138 individual observing programs up to $z \sim 8$ with median stellar mass $\sim 10^{8.4}$ to $10^{10}$

Arman Sauliere, Beatrice Magni, Guglielmo Lami, Xhek Turkeshi

We identify a \emph{universal functional form} that governs anticoncentration in random quantum circuits-one that holds across diverse circuit architectures and depths, and crucially remains valid even at finite system sizes and shallow depth. We support this claim through analytical results for ensembles of random tensor-network states and random-phase mode

Yuzhe Zhao, Congqiao Li, Antonios Agapitos, Dawei Fu

We present a novel method for measuring $|V_{cb}|$ at the LHC using an advanced boosted-jet tagger to identify "$bc$ signatures". By associating boosted $W \rightarrow bc$ signals with $bc$-matched jets from top-quark decays, we enable an in-situ calibration of the tagger. This approach significantly suppresses backgrounds while reducing uncertaintie

Alberto Colombo, Om Sharan Salafia, Giancarlo Ghirlanda, Francesco Iacovelli

The Einstein Telescope (ET), a proposed next-generation gravitational wave (GW) observatory, will expand the reach of GW astronomy of stellar-mass compact object binaries to unprecedented distances, enhancing opportunities for multi-messenger observations. Here we investigate multi-messenger emission properties of binary neutron star (NSNS) and black hole-ne

Aldana Grichener, Mathieu Renzo, Wolfgang E. Kerzendorf, Rob Farmer

One of the main challenges in modeling massive stars to the onset of core collapse is the computational bottleneck of nucleosynthesis during advanced burning stages. The number of isotopes formed requires solving a large set of fully-coupled stiff ordinary differential equations (ODEs), making the simulations computationally intensive and prone to numerical

Matthew A. Taylor, Behzad Tahmasebzadeh, Solveig Thompson, Eugene Vasiliev

The integral-field unit mode of the Near-Infrared Spectrograph (NIRSpec+IFU) mounted on the James Webb Space Telescope has now enabled kinematic studies of smaller and less massive compact stellar systems in which to search for central massive black holes (BHs) than ever before. We present here the first such detection using NIRSpec+IFU in its highest resolu

Gabriele Montefalcone, Gilly Elor, Kimberly K. Boddy, Nicola Bellomo

Leptophilic sub-MeV spin-zero dark matter (DM) decays into photons via one-loop processes, a scenario that has been in part overlooked in current literature. In this work, we provide updated and comprehensive upper limits on scalar, pseudo-scalar, and axion-like DM-electron couplings based on the latest NPIPE cosmic microwave background data from Planck. Our

Huanian Zhang, Guangping Ye, Rongyu Wu, Dennis Zaritsky

We present a machine learning search for local, low-mass galaxies ($z < 0.02$ and $10^6 M_\odot < M_* < 10^9 M_\odot$) using the combined photometric data from the DESI Imaging Legacy Surveys and the WISE survey. We introduce the spectrally confirmed training sample, discuss evaluation metrics, investigate the features, compare different machine learning alg

Subarna Chaki, Andrina Nicola, Alessio Spurio Mancini, Davide Piras

The simplest inflationary models predict the primordial power spectrum (PPS) of curvature perturbations to be nearly scale-invariant. However, various other models of inflation predict deviations from this behaviour, motivating a data-driven approach to reconstruct the PPS and constrain its shape. In this work, we present a novel method that employs a fully

William McClymont, Sandro Tacchella, Aaron Smith, Rahul Kannan

Characterizing the evolution of the star-forming main sequence (SFMS) at high redshift is crucial to contextualize the observed extreme properties of galaxies in the early Universe. We present an analysis of the SFMS and its scatter in the THESAN-ZOOM simulations, where we find a redshift evolution of the SFMS normalization scaling as $\propto (1+z)^{2.64\pm

Po-Shen Hsin, Ryohei Kobayashi, Carolyn Zhang

Anomalies of global symmetries provide important information on the quantum dynamics. We show the dynamical constraints can be organized into three classes: genuine anomalies, fractional topological responses, and integer responses that can be realized in symmetry-protected topological (SPT) phases. Coset symmetry can be present in many physical systems incl

Stephen J. McKay, Amy J. Barger, Lennox L. Cowie, Michael J. Nicandro Rosenthal

We identify a sample of 234 dusty star-forming galaxies (DSFGs) in the A2744 and GOODS-S fields using JWST/NIRCam-selected galaxies as priors for SCUBA-2 measurements. This method provides a large number of galaxies both above an 850$\,\mathrmμ$m flux of 2 mJy (47 bright DSFGs) and below (187 faint DSFGs), representing the largest sample of individually iden

Jan Boruch, Gabriele Di Ubaldo, Felix M. Haehl, Eric Perlmutter

We develop a non-perturbative definition of RMT${}_2$: a generalization of random matrix theory that is compatible with the symmetries of two-dimensional conformal field theory. Given any random matrix ensemble, its $n$-point spectral correlations admit a prescribed modular-invariant lift to RMT${}_2$, which moreover reduce to the original random matrix corr

Zheng Zhou

The Julia package FuzzifiED aims at simplifying the numerical calculations on the fuzzy sphere. It supports exact diagonalisation (ED) and density matrix renormalisation group (DMRG) calculations. FuzzifiED can also apply to generic fermionic and bosonic models. This documentation provides a review of the fuzzy sphere regularisation and an instruction for us

D. M. Williams, T. Treu, S. Birrer, A. J. Shajib

Gravitational time delays offer unique, independent measurements of the Hubble constant, $H_0$. Precise measurements of $H_0$ stand as one of the most pressing challenges in modern cosmology, and to do so with time delays requires precise lens models. While much work has focused on streamlining the modeling process to keep pace with the erumpent discovery of

Koutarou Kyutoku

We assess the universal relations among second-order moments of relativistic stars, namely the moment of inertia, tidal deformability, and spin-induced quadrupole moment, via reformulated perturbation equations. After constructing the spherical background configuration by solving two ordinary differential equations as usual, these three moments are obtained

Debasish Borah, Nayan Das, Nobuchika Okada, Prantik Sarmah

We study the possibility of the highest energy neutrino event with 220 PeV energy, detected recently by the KM3NeT experiment to be originating from heavy dark matter (DM) decay. Considering a heavy right handed neutrino (RHN) DM for illustrative purpose, we show that DM mass of 440 PeV, can explain the observed flux. The required DM lifetime to produce the

Komal Kumar, Tajamul Ashraf, Omkar Thawakar, Rao Muhammad Anwer

Large Language Models (LLMs) have transformed the natural language processing landscape and brought to life diverse applications. Pretraining on vast web-scale data has laid the foundation for these models, yet the research community is now increasingly shifting focus toward post-training techniques to achieve further breakthroughs. While pretraining provide

L. Degeorge, A. Ghosh, N. Dufour, D. Picard

Recent text-to-image (T2I) generation models have achieved remarkable sucess by training on billion-scale datasets, following a `bigger is better&#39; paradigm that prioritizes data quantity over availability (closed vs open source) and reproducibility (data decay vs established collections). We challenge this established paradigm by demonstrating that one c

Chiheb Ben Mahmoud, Zakariya El-Machachi, Krystian A. Gierczak, John L. A. Gardner

With the rapidly growing availability of machine-learned interatomic potential (MLIP) models for chemistry, much current research focuses on the development of generally applicable and ``foundational&#39;&#39; MLIPs. An important question in this context is whether, and how well, such models can transfer from one application domain to another. Here, we asses

Zhiyu Tan, Junyan Wang, Hao Yang, Luozheng Qin

Text-to-video generation has demonstrated promising progress with the advent of diffusion models, yet existing approaches are limited by dataset quality and computational resources. To address these limitations, this paper presents a comprehensive approach that advances both data curation and model design. We introduce CFC-VIDS-1M, a high-quality video datas

Chen-Kuan Lee

We construct stationary perturbations of a specific minimal surface with a circle of triple junctions in $\mathbb{R}^2 \times \mathbb{S}^1$, that satisfy given boundary data.

Yihong Dong, Ge Li, Xue Jiang, Yongding Tao

Periodicity, as one of the most important basic characteristics, lays the foundation for facilitating structured knowledge acquisition and systematic cognitive processes within human learning paradigms. However, the potential flaws of periodicity modeling in Transformer affect the learning efficiency and establishment of underlying principles from data for l

Thomas Waite, Yuang Geng, Trevor Turnquist, Ivan Ruchkin

It remains a challenge to provide safety guarantees for autonomous systems with neural perception and control. A typical approach obtains symbolic bounds on perception error (e.g., using conformal prediction) and performs verification under these bounds. However, these bounds can lead to drastic conservatism in the resulting end-to-end safety guarantee. This

Ludovico Mitchener, Jon M Laurent, Alex Andonian, Benjamin Tenmann

Large Language Models (LLMs) and LLM-based agents show great promise in accelerating scientific research. Existing benchmarks for measuring this potential and guiding future development continue to evolve from pure recall and rote knowledge tasks, towards more practical work such as literature review and experimental planning. Bioinformatics is a domain wher

Andrés Jaramillo Puentes, Roberto Pirisi

Using Galois-Stiefel-Whitney classes of theta characteristics we show that over a totally real base field the moduli stack of smooth genus $g$ curves and the moduli stack of principally polarized abelian varieties of dimension $g$ have nontrivial cohomological invariants and étale cohomology classes in degree respectively $2^{g-2}, 2^{g-1}$ and $2^{g-1}$. We

Yajnaseni Dutta, Lisa Marquand

A very good cubic fourfold is a smooth cubic fourfold that does not contain a plane, a cubic scroll, or a hyperplane section with a corank 3 singularity. We prove that the normalization of the relative compactified Prym variety associated to the universal family of hyperplanes of a very good cubic fourfold is in fact smooth, thereby extending prior results o

Vedran Brdar, Dibya S. Chattopadhyay

The KM3NeT collaboration recently reported the observation of KM3-230213A, a neutrino event with an energy exceeding 100 PeV, more than an order of magnitude higher than the most energetic neutrino in IceCube&#39;s catalog. Given its longer data-taking period and larger effective area relative to KM3NeT, IceCube should have observed events around that energy

Augusto Facundes, Kayman Jhosef Goncalves, Giorgio Torrieri

Motivated both by classical physics problems associated with ``Newton&#39;s bucket&#39;&#39; and recent developments related to QCD in rotating frames of reference relevant to heavy ion collisions, we discuss the difference between ``active&#39;&#39; and ``passive&#39;&#39; rotations in quantum systems. We examine some relevant potentials and give general sy

Xueyun Tian, Wei Li, Bingbing Xu, Yige Yuan

Despite significant progress in diffusion-based image generation, subject-driven generation and instruction-based editing remain challenging. Existing methods typically treat them separately, struggling with limited high-quality data and poor generalization. However, both tasks require capturing complex visual variations while maintaining consistency between

Pierre Houdouin, Lucas Saludjian

With the digitalization of power grids, physical equations become insufficient to describe the network&#39;s behavior, and realistic but time-consuming simulators must be used. Numerical experiments, such as safety validation, that involve simulating a large number of scenarios become computationally intractable. A popular solution to reduce the computationa

Hannah Sheats

In 1974, Erdős asked the following question: given a graph $G$ and a directed graph $\vec{H}$, how many ways are there to orient the edges of $G$ such that it does not contain $\vec{H}$ as a subgraph? We denote this value by $D(G, \vec{H})$. Further, we let $D(n, \vec{H})$ denote the maximum of $D(G, \vec{H})$ over all graphs $G$ on $n$ vertices. In 2006, Al

Laura Pierson

In arXiv:2301.02177, Crew, Pechenik, and Spirkl defined the Kromatic symmetric function $\overline{X}_G$ as a $K$-analogue of Stanley&#39;s chromatic symmetric function $X_G$, and one question they asked was how $\overline{X}_G$ expands in their $\overline{p}_λ$ basis, which they defined as a $K$-analogue of the classic power sum basis $p_λ.$ In arXiv:2408.0

Federico Di Gennaro, Thibault Laugel, Vincent Grari, Marcin Detyniecki

Traditional approaches to learning fair machine learning models often require rebuilding models from scratch, typically without considering potentially existing models. In a context where models need to be retrained frequently, this can lead to inconsistent model updates, as well as redundant and costly validation testing. To address this limitation, we intr

Michael E. Ressler, Alba Aller, David Jones, Ryan M. Lau

While NGC 1514 is an elliptical, but complex, planetary nebula at optical wavelengths, it was discovered to have a pair of infrared-bright, axisymmetric rings contained within its faint outer shell during the course of the WISE all-sky survey. We have obtained JWST mid-infrared imaging and spectroscopy of the nebula through the use of simultaneous observatio

Sherlon Almeida da Silva, Davi Geiger, Luiz Velho, Moacir Antonelli Ponti

We innovate in stereo vision by explicitly providing analytical 3D surface models as viewed by a cyclopean eye model that incorporate depth discontinuities and occlusions. This geometrical foundation combined with learned stereo features allows our system to benefit from the strengths of both approaches. We also invoke a prior monocular model of surfaces to

Marius F. R. Juston, William R. Norris, Dustin Nottage, Ahmet Soylemezoglu

Deep residual networks (ResNets) have demonstrated outstanding success in computer vision tasks, attributed to their ability to maintain gradient flow through deep architectures. Simultaneously, controlling the Lipschitz bound in neural networks has emerged as an essential area of research for enhancing adversarial robustness and network certifiability. This

Kulin Shah, Alkis Kalavasis, Adam R. Klivans, Giannis Daras

There is strong empirical evidence that the state-of-the-art diffusion modeling paradigm leads to models that memorize the training set, especially when the training set is small. Prior methods to mitigate the memorization problem often lead to a decrease in image quality. Is it possible to obtain strong and creative generative models, i.e., models that achi

Yuan Bian, Grace Y. Yi, Wenqing He

Boosting has emerged as a useful machine learning technique over the past three decades, attracting increased attention. Most advancements in this area, however, have primarily focused on numerical implementation procedures, often lacking rigorous theoretical justifications. Moreover, these approaches are generally designed for datasets with fully observed d