Tootfinder

Graph-based Integrated Gradients for Explaining Graph Neural Networks
Lachlan Simpson, Kyle Millar, Adriel Cheng, Cheng-Chew Lim, Hong Gunn Chew
https://arxiv.org/abs/2509.07648

Graph-based Integrated Gradients for Explaining Graph Neural Networks
Integrated Gradients (IG) is a common explainability technique to address the black-box problem of neural networks. Integrated gradients assumes continuous data. Graphs are discrete structures making IG ill-suited to graphs. In this work, we introduce graph-based integrated gradients (GB-IG); an extension of IG to graphs. We demonstrate on four synthetic datasets that GB-IG accurately identifies crucial structural components of the graph used in classification tasks. We further demonstrate on t…

Pathological solutions of Navier-Stokes equations on $\mathbb{T}^2$ with gradients in Hardy spaces
Jan Burczak, Antonio Hidalgo-Torn\'e
https://arxiv.org/abs/2509.08168 http…

Pathological solutions of Navier-Stokes equations on $\mathbb{T}^2$ with gradients in Hardy spaces
For an arbitrary smooth initial datum, we construct multiple nonzero solutions to the $2$d Navier-Stokes equations, with their gradients in the Hardy space $\mathcal{H}^p$ with any $p \in (0,1)$. Thus, in terms of the path space $C(\mathcal{H}^p)$ for vorticity, $p=1$ is the threshold value distinguishing between non-uniqueness and uniqueness regimes. In order to obtain our result, we develop the needed theory of Hardy spaces on periodic domains.

IDK; I'm autistic, and I think queer and bisexual people are cool! 🏳️‍🌈
#ActuallyAutistic #Autism #Autistic

Stochastic resolution of identity to CC2 for large systems: Excited-state gradients and derivative couplings
Chongxiao Zhao, Chenyang Li, Wenjie Dou
https://arxiv.org/abs/2509.06460

Stochastic resolution of identity to CC2 for large systems: Excited-state gradients and derivative couplings
Excited-state gradients and derivative couplings are critical for simulating excited-state dynamics. However, their calculations are very expensive within the coupled-cluster framework due to the steep scaling. In this work, we present two implementations of stochastic resolution of identity to CC2 (sRI-CC2) for excited-state analytical gradients and derivative couplings. The first method employs sRI for both Coulomb and exchange terms, reducing the formal scaling to cubic. However, this method…

Adaptive collaboration for online personalized distributed learning with heterogeneous clients
Constantin Philippenko, Batiste Le Bars, Kevin Scaman, Laurent Massouli\'e
https://arxiv.org/abs/2507.06844

Adaptive collaboration for online personalized distributed learning with heterogeneous clients
We study the problem of online personalized decentralized learning with $N$ statistically heterogeneous clients collaborating to accelerate local training. An important challenge in this setting is to select relevant collaborators to reduce gradient variance while mitigating the introduced bias. To tackle this, we introduce a gradient-based collaboration criterion, allowing each client to dynamically select peers with similar gradients during the optimization process. Our criterion is motivated…

Ampere: Communication-Efficient and High-Accuracy Split Federated Learning
Zihan Zhang, Leon Wong, Blesson Varghese
https://arxiv.org/abs/2507.07130 https:…

Ampere: Communication-Efficient and High-Accuracy Split Federated Learning
A Federated Learning (FL) system collaboratively trains neural networks across devices and a server but is limited by significant on-device computation costs. Split Federated Learning (SFL) systems mitigate this by offloading a block of layers of the network from the device to a server. However, in doing so, it introduces large communication overheads due to frequent exchanges of intermediate activations and gradients between devices and the server and reduces model accuracy for non-IID data. W…

An Energy Perspective of Core Erosion in Gas Giant Planets
J. R. Fuentes, Christopher R. Mankovich, Ankan Sur
https://arxiv.org/abs/2507.05109 https://

An Energy Perspective of Core Erosion in Gas Giant Planets
Juno and Cassini have shown that Jupiter and Saturn likely contain extended gradients of heavy elements. Yet, how these gradients can survive over billions of years remains an open question. Classical convection theories predict rapid mixing and homogenization, which would erase such gradients on timescales far shorter than the planets' ages. To address this, we estimate the energy required to erode both dense and fuzzy cores, and compare it to what the planet can realistically supply. If the e…

Breaking the Conventional Forward-Backward Tie in Neural Networks: Activation Functions
Luigi Troiano, Francesco Gissi, Vincenzo Benedetto, Genny Tortora
https://arxiv.org/abs/2509.07236

Breaking the Conventional Forward-Backward Tie in Neural Networks: Activation Functions
Gradient-based neural network training traditionally enforces symmetry between forward and backward propagation, requiring activation functions to be differentiable (or sub-differentiable) and strictly monotonic in certain regions to prevent flat gradient areas. This symmetry, linking forward activations closely to backward gradients, significantly restricts the selection of activation functions, particularly excluding those with substantial flat or non-differentiable regions. In this paper, we…

Replaced article(s) found for cond-mat.stat-mech. https://arxiv.org/list/cond-mat.stat-mech/new
[1/1]:
- Learning to control non-equilibrium dynamics using local imperfect gradients
Carlos Floyd, Aaron R. Dinner, Suriyanarayanan Vaikuntanathan

Decoupling pressure gradient history effects in turbulent boundary layers through high-Reynolds number experiments
Ahmad Zarei, Mitchell Lozier, Rahul Deshpande, Ivan Marusic
https://arxiv.org/abs/2509.07545

Decoupling pressure gradient history effects in turbulent boundary layers through high-Reynolds number experiments
This study provides a carefully controlled examination of the universality of the von Karman and additive constants associated with the classical logarithmic scaling of the mean streamwise velocity profile in high-friction Reynolds number (Re_tau) turbulent boundary layers (TBLs) subjected to weak-to-moderate adverse pressure gradients (APGs). The analysis leverages a recently developed method for imposing APGs with minimal pressure gradient (PG) history effects in Melbourne's high-Re_tau TBL f…

Multi-module GRPO: Composing Policy Gradients and Prompt Optimization for Language Model Programs
Noah Ziems, Dilara Soylu, Lakshya A Agrawal, Isaac Miller, Liheng Lai, Chen Qian, Kaiqiang Song, Meng Jiang, Dan Klein, Matei Zaharia, Karel D'Oosterlinck, Christopher Potts, Omar Khattab
https://arxiv.org/abs/2508.04660

Multi-module GRPO: Composing Policy Gradients and Prompt Optimization for Language Model Programs
Group Relative Policy Optimization (GRPO) has proven to be an effective tool for post-training language models (LMs). However, AI systems are increasingly expressed as modular programs that mix together multiple LM calls with distinct prompt templates and other tools, and it is not clear how best to leverage GRPO to improve these systems. We begin to address this challenge by defining mmGRPO, a simple multi-module generalization of GRPO that groups LM calls by module across rollouts and handles…

Universal Representation of Generalized Convex Functions and their Gradients
Moeen Nehzati
https://arxiv.org/abs/2509.04477 https://arxiv.org/pdf/2509.0447…

Universal Representation of Generalized Convex Functions and their Gradients
Solutions to a wide range of optimization problems, from optimal transport theory to mathematical economics, often take the form of generalized convex functions (GCFs). This characterization can be used to convert nested bilevel optimization problems into single-level optimization problems. Despite this, the characterization has not been fully exploited in numerical optimization. When the solution to an optimization problem is known to belong to a particular class of objects, this information…

The Open Cluster Chemical Abundances and Mapping Survey: VIII. Galactic Chemical Gradient and Azimuthal Analysis from SDSS/MWM DR19
Jonah M. Otto, Peter M Frinchaboy, Natalie R. Myers, James W. Johnson, John Donor, Ahabar Hossain, Szabolcs M\'esz\'aros, Katia Cunha, Binod Bhattarai, Gail Zasowski, Sarah R. Loebman, Alessa I. Wiggins, Adrian M. Price-Whelan, Taylor Spoo, Diogo Souto, Dmitry Bizyaev, Kaike Pan, Andrew K. Saydjari

The Open Cluster Chemical Abundances and Mapping Survey: VIII. Galactic Chemical Gradient and Azimuthal Analysis from SDSS/MWM DR19
The Open Cluster Chemical Abundances and Mapping (OCCAM) survey seeks to curate a large, comprehensive, uniform dataset of open clusters and member stars to constrain key Galactic parameters. This eighth entry from the OCCAM survey, based on the newly released SDSS-V/MWM Data Release 19 (DR19), has established a sample of 164 high quality open clusters that are used to constrain the radial and azimuthal gradients of the Milky Way. The DR19 cluster sample [Fe/H] abundances are roughly consistent…

Meridional circulation molecular-weighted
Deepayan Banik, Kristen Menou, Evan H. Anders
https://arxiv.org/abs/2508.03844 https://arxiv.org/pdf/2508.03844…

Meridional circulation molecular-weighted
Meridional circulation in stratified stellar/planetary interiors in the presence of stable molecular weight gradients remains poorly understood, thereby affecting angular momentum transport in evolutionary models. We extend the downward control principle of atmospheric sciences to include compositional stratification. Using a linearized analysis we show that stable compositional gradients slow down the penetration of circulation into the depths, emphasizing the importance of time-dependent solu…

Time-resolved measurement of Seebeck effect for superionic metals during structural phase transition
Shilin Li (Beijing Academy of Quantum Information Sciences), Hailiang Xia (Beijing Huihaoyu Intelligent Technology Co., Ltd), Takuma Ogasawara (Beijing Academy of Quantum Information Sciences), Liguo Zhang (Beijing Academy of Quantum Information Sciences), Katsumi Tanigaki (Beijing Academy of Quantum Information Sciences)

Time-resolved measurement of Seebeck effect for superionic metals during structural phase transition
We propose a new time (t)-resolved method of both vertical- and horizontal-temperature gradients in an orthogonal configuration (t-resolved T(t)-HVOT) to have real interpretations of the enhancement in thermoelectric Seebeck effect (SE) observed during the structural phase transition. We apply our new method to superionic-state semiconductors of p-type Cu2Se and n-type Ag2S. The experimental data differentiate the two types of enhancements during the phase transition: a colossal SE (Scolossal),…

Programming tension in 3D printed networks inspired by spiderwebs
Thijs Masmeijer, Caleb Swain, Jeff Hill, Ed Habtour
https://arxiv.org/abs/2509.05855 https://

Programming tension in 3D printed networks inspired by spiderwebs
Each element in tensioned structural networks -- such as tensegrity, architectural fabrics, or medical braces/meshes -- requires a specific tension level to achieve and maintain the desired shape, stability, and compliance. These structures are challenging to manufacture, 3D print, or assemble because flattening the network during fabrication introduces multiplicative inaccuracies in the network's final tension gradients. This study overcomes this challenge by offering a fabrication algorithm f…

FACT: the Features At Convergence Theorem for neural networks
Enric Boix-Adsera, Neil Mallinar, James B. Simon, Mikhail Belkin
https://arxiv.org/abs/2507.05644

FACT: the Features At Convergence Theorem for neural networks
A central challenge in deep learning theory is to understand how neural networks learn and represent features. To this end, we prove the Features at Convergence Theorem (FACT), which gives a self-consistency equation that neural network weights satisfy at convergence when trained with nonzero weight decay. For each weight matrix $W$, this equation relates the "feature matrix" $W^\top W$ to the set of input vectors passed into the matrix during forward propagation and the loss gradients passed t…

Spatially localised doubly diffusive convection in an axisymmetric spherical shell
Paul M. Mannix, Cedric Beaume
https://arxiv.org/abs/2507.03830 https://

Spatially localised doubly diffusive convection in an axisymmetric spherical shell
Doubly diffusive convection describes the fluid motion driven by the competition of temperature and salinity gradients diffusing at different rates. While the convective motions driven by these gradients usually occupy the entire domain, parameter regions exist where the convection is spatially localised. Although well-studied in planar geometries, spatially localised doubly diffusive convection has never been investigated in a spherical shell, a geometry of relevance to astrophysics. In this p…

The Evolution and Internal Structure of Neptunes and Sub-Neptunes: The importance of thermal conductivity in non-convective regions
Mark Eberlein, Ravit Helled
https://arxiv.org/abs/2509.04564

The Evolution and Internal Structure of Neptunes and Sub-Neptunes: The importance of thermal conductivity in non-convective regions
Neptunes and sub-Neptunes are typically modeled under the assumption that the interior is adiabatic and consists of distinct layers. However, formation models indicate that composition gradients can exist. Such composition gradients can significantly affect the planetary thermal evolution. In non-convective layers, the heat transport is governed by multiple processes. We investigate how the evolution and internal structure of Neptunes and sub-Neptunes is affected when considering non-convective…

On the Inherent Privacy of Zeroth Order Projected Gradient Descent
Devansh Gupta, Meisam Razaviyayn, Vatsal Sharan
https://arxiv.org/abs/2507.05610 https:/…

On the Inherent Privacy of Zeroth Order Projected Gradient Descent
Differentially private zeroth-order optimization methods have recently gained popularity in private fine tuning of machine learning models due to their reduced memory requirements. Current approaches for privatizing zeroth-order methods rely on adding Gaussian noise to the estimated zeroth-order gradients. However, since the search direction in the zeroth-order methods is inherently random, researchers including Tang et al. (2024) and Zhang et al. (2024a) have raised an important question: is t…

Fully Analytic Nuclear Gradients for the Bethe--Salpeter Equation
Johannes T\"olle, Marios-Petros Kitsaras, Pierre-Fran\c{c}ois Loos
https://arxiv.org/abs/2507.02160

Fully Analytic Nuclear Gradients for the Bethe--Salpeter Equation
The Bethe-Salpeter equation (BSE) formalism, combined with the $GW$ approximation for ionization energies and electron affinities, is emerging as an efficient and accurate method for predicting optical excitations in molecules. In this letter, we present the first derivation and implementation of fully analytic nuclear gradients for the BSE@$G_0W_0$ method. Building on recent developments for $G_0W_0$ nuclear gradients, we derive analytic nuclear gradients for several BSE@$G_0W_0$ variants. We …

Midveins regulate the shape formation of drying leaves
Kexin Guo, Yafei Zhang, Massimo Paradiso, Yuchen Long, K. Jimmy Hsia, Mingchao Liu
https://arxiv.org/abs/2507.01813

Midveins regulate the shape formation of drying leaves
Dried leaves in nature often exhibit curled and crumpled morphologies, typically attributed to internal strain gradients that produce dome-like shapes. However, the origin of these strain gradients remains poorly understood. Although leaf veins--particularly the midvein--have been suggested to influence shape formation, their mechanical role has not been systematically investigated. Here, we demonstrate that mechanical constraints imposed by the midvein play a crucial role in generating the div…

Universal Scaling Formalism and Analytical Optimization Criterion for Multiscale Geometric Design of Thermoelectric Metamaterials
Xanthippi Zianni
https://arxiv.org/abs/2509.05095

Universal Scaling Formalism and Analytical Optimization Criterion for Multiscale Geometric Design of Thermoelectric Metamaterials
Thermoelectric (TE) generators can directly convert heat into electricity, but their performance is often constrained by limited temperature gradients. Here it is shown that width-modulated metamaterials with constrictions and expansions (constricted geometries) enhance temperature difference DT by reduced Transmissivity (Tr), a geometry-based parameter defined by the ratio of constriction to expansion cross-sections. A universal scaling behavior of transport and key TE efficiency metrics with …

Unified Statistical Theory of Heat Conduction in Nonuniform Media
Yi Zeng, Jianjun Dong
https://arxiv.org/abs/2507.05682 https://arxi…

Unified Statistical Theory of Heat Conduction in Nonuniform Media
We present a theory of heat conduction that encompasses and extends the classical paradigm defined by Fourier law and the Kapitza interfacial conductance model. Derived from the Zwanzig projection operator formalism, our approach introduces a spatiotemporal kernel function of heat conduction, which governs the spatiotemporal response of the local heat flux in nonuniform media to surrounding temperature gradients while incorporating temporal memory effects. Our statistical analysis shows that me…

The Light Echo of a High-Redshift Quasar mapped with Lyman-$\alpha$ Tomography
Anna-Christina Eilers, Minghao Yue, Jorryt Matthee, Joseph F. Hennawi, Frederick B. Davies, Robert A. Simcoe, Richard Teague, Rongmon Bordoloi, Gabriel Brammer, Yi Kang, Daichi Kashino, Ruari Mackenzie, Rohan P. Naidu, Benjamin Navarrete
https://arxiv.org/abs/25…

The Light Echo of a High-Redshift Quasar mapped with Lyman-$α$ Tomography
Ultra-violet (UV) radiation from accreting black holes ionizes the intergalactic gas around early quasars, carving out highly ionized bubbles in their surroundings. Any changes in a quasar's luminosity are therefore predicted to produce outward-propagating ionization gradients, affecting the Lyman-$α$ (Ly$α$) absorption opacity near the quasar's systemic redshift. This "proximity effect" is well-documented in rest-UV quasar spectra but only provides a one-dimensional probe along our line-of-s…

Just learned that my little visual cosine gradient generator tool (from 2015) is featured in this popular video tutorial (1.1M views) about artistic shader programming:
https://youtu.be/f4s1h2YETNY?t=971
The tool is still here:

Efficient GPU-Accelerated Training of a Neuroevolution Potential with Analytical Gradients
Hongfu Huang, Junhao Peng, Kaiqi Li, Jian Zhou, Zhimei Sun
https://arxiv.org/abs/2507.00528

Efficient GPU-Accelerated Training of a Neuroevolution Potential with Analytical Gradients
Machine-learning interatomic potentials (MLIPs) such as neuroevolution potentials (NEP) combine quantum-mechanical accuracy with computational efficiency significantly accelerate atomistic dynamic simulations. Trained by derivative-free optimization, the normal NEP achieves good accuracy, but suffers from inefficiency due to the high-dimensional parameter search. To overcome this problem, we present a gradient-optimized NEP (GNEP) training framework employing explicit analytical gradients and t…

Adjoint-Based Aerodynamic Shape Optimization with a Manifold Constraint Learned by Diffusion Models
Long Chen, Emre Oezkaya, Jan Rottmayer, Nicolas R. Gauger, Zebang Shen, Yinyu Ye
https://arxiv.org/abs/2507.23443

Adjoint-Based Aerodynamic Shape Optimization with a Manifold Constraint Learned by Diffusion Models
We introduce an adjoint-based aerodynamic shape optimization framework that integrates a diffusion model trained on existing designs to learn a smooth manifold of aerodynamically viable shapes. This manifold is enforced as an equality constraint to the shape optimization problem. Central to our method is the computation of adjoint gradients of the design objectives (e.g., drag and lift) with respect to the manifold space. These gradients are derived by first computing shape derivatives with res…

Competing chemical gradients change chemotactic dynamics and cell distribution
Emiliano Perez Ipi\~na, Brian A. Camley
https://arxiv.org/abs/2507.19341 https://

Competing chemical gradients change chemotactic dynamics and cell distribution
Cells are constantly exposed to diverse stimuli-chemical, mechanical, or electrical-that guide their movement. In physiological conditions, these signals often overlap, as seen during infections, where neutrophils and dendritic cells navigate through multiple chemotactic fields. How cells integrate and prioritize competing signals remains unclear. For instance, in the presence of opposing chemoattractant gradients, how do cells decide which direction to go? When should local signals dominate di…

Hard Contacts with Soft Gradients: Refining Differentiable Simulators for Learning and Control
Anselm Paulus, A. Ren\'e Geist, Pierre Schumacher, V\'it Musil, Georg Martius
https://arxiv.org/abs/2506.14186

Hard Contacts with Soft Gradients: Refining Differentiable Simulators for Learning and Control
Contact forces pose a major challenge for gradient-based optimization of robot dynamics as they introduce jumps in the system's velocities. Penalty-based simulators, such as MuJoCo, simplify gradient computation by softening the contact forces. However, realistically simulating hard contacts requires very stiff contact settings, which leads to incorrect gradients when using automatic differentiation. On the other hand, using non-stiff settings strongly increases the sim-to-real gap. We analyze …

A Generalized $\ell_1$-Merit Function SQP Method Using Function Approximations with Tunable Accuracy
Dane S. Grundvig, Matthias Heinkenschloss
https://arxiv.org/abs/2507.06199

A Generalized $\ell_1$-Merit Function SQP Method Using Function Approximations with Tunable Accuracy
This paper develops a generalization of the line-search sequential quadratic programming (SQP) algorithm with $\ell_1$-merit function that uses objective and constraint function approximations with tunable accuracy to solve smooth equality-constrained optimization problems. The evaluation of objective and constraint functions and their gradients is potentially computationally expensive, but it is assumed that one can construct effective, computationally inexpensive models of these functions. Th…

Measuring high field gradients of cobalt nanomagnets in a spin-mechanical setup
Felix Hahne, Teresa Klara Pfau, Liza \v{Z}aper, Lucio Stefan, Thibault Capelle, Andrea Ranfagni, Martino Poggio, Albert Schliesser
https://arxiv.org/abs/2508.19156

Measuring high field gradients of cobalt nanomagnets in a spin-mechanical setup
Hybrid systems composed of a single nitrogen-vacancy center spin magnetically coupled to a macroscopic mechanical resonator constitute promising platforms for the realization of quantum information protocols and for quantum sensing applications. The magnetic structure that mediates the interaction must ensure high field gradients while preserving the spin and mechanical properties. We present a spin-mechanical setup built around a cobalt nanomagnet grown with focused electron beam-induced depos…

Replaced article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[4/5]:
- Avoid Forgetting by Preserving Global Knowledge Gradients in Federated Learning with Non-IID Data
Abhijit Chunduru, Majid Morafah, Mahdi Morafah, Vishnu Pandi Chellapandi, Ang Li

Equivalence of Koopman Eigenfunctions and a Commuting Local Frame of Symmetries
Xinyuan Jiang, Yan Li
https://arxiv.org/abs/2508.02437 https://arxiv.org/pd…

Equivalence of Koopman Eigenfunctions and a Commuting Local Frame of Symmetries
This article establishes the relationship between the Koopman eigenfunctions of a first-order nonlinear ODE system and its symmetries. Specifically, a bijective mapping is obtained between i) a commuting local frame of symmetry generators, and ii) a set of Koopman eigenfunctions whose gradients form a local frame. This equivalence is used to convert the problem of finding Koopman eigenfunctions into the equivalent problem of finding commuting vector fields, which are readily given by the flow m…

Single Epoch Measurements of Dispersion Measure Gradients Towards PSR B0834 06
Daniel Baker, Ue-Li Pen
https://arxiv.org/abs/2508.15270 https://arxiv.org/p…

Single Epoch Measurements of Dispersion Measure Gradients Towards PSR B0834+06
Understanding the evolution of pulsar dispersion measures is vital to high precision timing experiments, as well as astrometric experiments to determine pulsar positions and proper motions. In this work, we present a novel approach to measuring dispersion measure gradients using pulsar scintillometry. This approach makes use of the multipath propagation through the interstellar medium to simultaneously probe dispersion measures along multiple sight lines from a single observation. Using existin…

Maximal Double-Exponential Growth for the Euler Equation on the Half-Plane
Andrej Zlatos
https://arxiv.org/abs/2507.04198 https://arx…

Maximal Double-Exponential Growth for the Euler Equation on the Half-Plane
We show that smooth solutions to the Euler equation on the half-plane can exhibit double-exponential growth of their vorticity gradients. We also determine the maximal possible growth rate and construct solutions that saturate it. These appear to be the first such results on an unbounded resp. any 2D domain.

Analysis on asymmetric metric measure spaces: $q$-heat flow, $q$-Laplacian and Sobolev spaces
Alexandru Krist\'aly, Shin-ichi Ohta, Wei Zhao
https://arxiv.org/abs/2509.03268

Analysis on asymmetric metric measure spaces: $q$-heat flow, $q$-Laplacian and Sobolev spaces
We investigate geometric analysis on metric measure spaces equipped with asymmetric distance functions. Extending concepts from the symmetric case, we introduce upper gradients and corresponding $L^q$-energy functionals as well as $q$-Laplacian in the asymmetric setting. Along the lines of gradient flow theory, we then study $q$-heat flow as the $L^2$-gradient flow for $L^q$-energy. We also study the Sobolev spaces over asymmetric metric measure spaces, extending some results of Ambrosio--Gigli…

TUC-PPO: Team Utility-Constrained Proximal Policy Optimization for Spatial Public Goods Games
Zhaoqilin Yang, Xin Wang, Ruichen Zhang, Chanchan Li, Youliang Tian
https://arxiv.org/abs/2507.02675

TUC-PPO: Team Utility-Constrained Proximal Policy Optimization for Spatial Public Goods Games
We introduce Team Utility-Constrained Proximal Policy Optimization (TUC-PPO), a new deep reinforcement learning framework. It extends Proximal Policy Optimization (PPO) by integrating team welfare objectives specifically for spatial public goods games. Unlike conventional approaches where cooperation emerges indirectly from individual rewards, TUC-PPO instead optimizes a bi-level objective integrating policy gradients and team utility constraints. Consequently, all policy updates explicitly inc…

Gradients of unitary optical neural networks using parameter-shift rule
Jinzhe Jiang, Yaqian Zhao, Xin Zhang, Chen Li, Yunlong Yu, Hailing Liu
https://arxiv.org/abs/2506.11565

Gradients of unitary optical neural networks using parameter-shift rule
This paper explores the application of the parameter-shift rule (PSR) for computing gradients in unitary optical neural networks (UONNs). While backpropagation has been fundamental to training conventional neural networks, its implementation in optical neural networks faces significant challenges due to the physical constraints of optical systems. We demonstrate how PSR, which calculates gradients by evaluating functions at shifted parameter values, can be effectively adapted for training UONNs…

Deep Feature-specific Imaging
Yizhou Lu
https://arxiv.org/abs/2508.01981 https://arxiv.org/pdf/2508.01981

Deep Feature-specific Imaging
Modern photon-counting sensors are increasingly dominated by Poisson noise, yet conventional Feature-Specific Imaging (FSI) is optimized for additive Gaussian noise, leading to suboptimal performance and a loss of its advantages under Poisson noise. To address this, we introduce DeepFSI, a novel end-to-end optical-electronic framework. DeepFSI "unfreezes" traditional FSI masks, enabling a deep neural network to learn globally optimal measurement masks by computing gradients directly under reali…

Numerical simulations of electron acceleration driven by heavy ion beams in plasma with alternating density gradients
Jiangdong Li, Jiancheng Yang, Guoxing Xia, Jie Liu, Ruihu Zhu, Xiangwen Qiao
https://arxiv.org/abs/2507.02269

Numerical simulations of electron acceleration driven by heavy ion beams in plasma with alternating density gradients
Plasma-Based Acceleration (PBA) has been demonstrated using laser, electron, and proton drivers. However, significant challenges remain in achieving high efficiency, stable acceleration, and scalable energy gain. Heavy ion beam drivers, with their high kinetic energy, offer the potential for greater energy transfer to the witness beam. Unfortunately, limited by the relatively low velocity of heavy ion, the dephasing length is really short leading to a low energy gain of the witness beam. Conven…

A Practical and Secure Byzantine Robust Aggregator
De Zhang Lee, Aashish Kolluri, Prateek Saxena, Ee-Chien Chang
https://arxiv.org/abs/2506.23183 https://

A Practical and Secure Byzantine Robust Aggregator
In machine learning security, one is often faced with the problem of removing outliers from a given set of high-dimensional vectors when computing their average. For example, many variants of data poisoning attacks produce gradient vectors during training that are outliers in the distribution of clean gradients, which bias the computed average used to derive the ML model. Filtering them out before averaging serves as a generic defense strategy. Byzantine robust aggregation is an algorithmic pri…

Zinnias are another favorite of mine that I'm stocking up on, the best thing about these is they come in such a wide range of colors and gradients of color
#photo #photography #flowers

A Brenier Theorem on $(\mathcal{P}_2 (\mathcal{P}_2(\mathbb{R}^d )), W_2 )$ and Applications to Adapted Transport
Mathias Beiglb\"ock, Gudmund Pammer, Stefan Schrott
https://arxiv.org/abs/2509.03506

A Brenier Theorem on $(\mathcal{P}_2 (\mathcal{P}_2(\mathbb{R}^d )), W_2 )$ and Applications to Adapted Transport
Brenier's fundamental theorem characterizes optimal transport plans for measures $μ, ν$ on $\mathbb{R}^d$ and quadratic distance costs in terms of gradients of convex functions. In particular it guarantees the existence of optimal transport maps for measures which are absolutely continuous wrt Lebesgue measure. Our goal is to provide a version of this result for measures $P,Q$ on $\mathcal{P}_2(\mathbb{R}^d)$ and costs given by the squared Wasserstein distance $W_2^2(μ, ν)$. We characteri…

Stochastic Gradients under Nuisances
Facheng Yu, Ronak Mehta, Alex Luedtke, Zaid Harchaoui
https://arxiv.org/abs/2508.20326 https://arxiv.org/pdf/2508.2032…

Stochastic Gradients under Nuisances
Stochastic gradient optimization is the dominant learning paradigm for a variety of scenarios, from classical supervised learning to modern self-supervised learning. We consider stochastic gradient algorithms for learning problems whose objectives rely on unknown nuisance parameters, and establish non-asymptotic convergence guarantees. Our results show that, while the presence of a nuisance can alter the optimum and upset the optimization trajectory, the classical stochastic gradient algorithm …

What if each voxel were measured with a different diffusion protocol?
Santiago Coelho, Gregory Lemberskiy, Ante Zhu, Hong-Hsi Lee, Nastaren Abad, Thomas K. F. Foo, Els Fieremans, Dmitry S. Novikov
https://arxiv.org/abs/2506.22650

What if each voxel were measured with a different diffusion protocol?
Expansion of diffusion MRI (dMRI) both into the realm of strong gradients, and into accessible imaging with portable low-field devices, brings about the challenge of gradient nonlinearities. Spatial variations of the diffusion gradients make diffusion weightings and directions non-uniform across the field of view, and deform perfect shells in the q-space designed for isotropic directional coverage. Such imperfections hinder parameter estimation: Anisotropic shells hamper the deconvolution of fi…

Diffusion model for gradient preconditioning in hyperspectral imaging inverse problems
Jonathan Monsalve, Kumar Vijay Mishra
https://arxiv.org/abs/2507.23065 https://

Diffusion model for gradient preconditioning in hyperspectral imaging inverse problems
Recovering high-dimensional statistical structure from limited measurements is a fundamental challenge in hyperspectral imaging, where capturing full-resolution data is often infeasible due to sensor, bandwidth, or acquisition constraints. A common workaround is to partition measurements and estimate local statistics-such as the covariance matrix-using only partial observations. However, this strategy introduces noise in the optimization gradients, especially when each partition contains few sa…

Novel Knockoff Generation and Importance Measures with Heterogeneous Data via Conditional Residuals and Local Gradients
Evan Mason, Zhe Fei
https://arxiv.org/abs/2508.14882 http…

Novel Knockoff Generation and Importance Measures with Heterogeneous Data via Conditional Residuals and Local Gradients
Knockoff variable selection is a powerful framework that creates synthetic knockoff variables to mirror the correlation structure of the observed features, enabling principled control of the false discovery rate in variable selection. However, existing methods often assume homogeneous data types or known distributions, limiting their applicability in real-world settings with heterogeneous, distribution-free data. Moreover, common variable importance measures rely on linear outcome models, hinde…

Asymptotically Optimal Secure Aggregation for Wireless Federated Learning with Multiple Servers
Zhenhao Huang, Kai Liang, Yuanming Shi, Songze Li, Youlong Wu
https://arxiv.org/abs/2506.23680

Asymptotically Optimal Secure Aggregation for Wireless Federated Learning with Multiple Servers
In this paper, we investigate the transmission latency of the secure aggregation problem in a \emph{wireless} federated learning system with multiple curious servers. We propose a privacy-preserving coded aggregation scheme where the servers can not infer any information about the distributed users' local gradients, nor the aggregation value. In our scheme, each user encodes its local gradient into $\sK$ confidential messages intended exclusively for different servers using a multi-secret shari…

Anisotropic spin fluctuations in the triangular Kondo lattice compound CePtAl$_4$Ge$_2$ probed by site-selective $^{27}$Al NMR
H. Sakai, S. Shin, S. Kambe, Y. Tokunaga, H. Harima, E. Pomjakushina, T. Park
https://arxiv.org/abs/2509.03007

Anisotropic spin fluctuations in the triangular Kondo lattice compound CePtAl$_4$Ge$_2$ probed by site-selective $^{27}$Al NMR
A site-selective $^{27}$Al nuclear magnetic resonance (NMR) study is carried out on the Kondo lattice compound CePtAl$_4$Ge$_2$, which crystallizes in a rhombohedral lattice with quasi-two-dimensional Ce layers forming a triangular lattice network. Two inequivalent Al sites, Al(1) and Al(2), are unambiguously assigned by comparing measured nuclear quadrupole parameters with electric field gradients obtained from electronic structure calculations. Knight shift analysis yields distinct hyperfine …

Out-of-time-order correlators bridge classical transport and quantum dynamics
Sophia N. Fricke, Haiyan Mao, Manas Sajjan, Ashok Ajoy, Velencia Witherspoon, Sabre Kais, Jeffrey A. Reimer
https://arxiv.org/abs/2508.20235

Out-of-time-order correlators bridge classical transport and quantum dynamics
The out-of-time-order correlator (OTOC) has emerged as a central tool for quantifying decoherence across wide-ranging physical platforms. Here we demonstrate its direct measurement in a classical ensemble using nuclear magnetic resonance (NMR) with a modulated gradient spin echo (MGSE) sequence and extend the method into a multidimensional correlation to track exchange phenomena. Position is encoded through magnetic field gradients and momentum through the velocity autocorrelation function, ena…

Nuclear Stellar Discs
Mathias Schultheis (Universit\'e C\^ote d'Azur), Mattia C. Sormani (Universit\'a dell'Insubria), Dimitri A. Gadotti (Durham University)
https://arxiv.org/abs/2509.04562

Nuclear Stellar Discs
We review our current understanding of nuclear stellar discs (NSDs), rotating, and flattened stellar structures found in the central regions of both early- and late-type galaxies. We examine their demographics, kinematics, stellar populations, metallicity gradients and star formation histories. We derive scaling relations linking NSDs to properties of their host galaxies, and compare them with analogous relations for nuclear star clusters. The relationship between NSDs and other central galacti…

Protein Drift-Diffusion in Membranes with Non-equilibrium Fluctuations arising from Gradients in Concentration or Temperature
D. Jasuja, P. J. Atzberger
https://arxiv.org/abs/2506.22695

Protein Drift-Diffusion in Membranes with Non-equilibrium Fluctuations arising from Gradients in Concentration or Temperature
We investigate proteins within heterogeneous cell membranes where non-equilibrium phenomena arises from spatial variations in concentration and temperature. We develop simulation methods building on non-equilibrium statistical mechanics to obtain stochastic hybrid continuum-discrete descriptions which track individual protein dynamics, spatially varying concentration fluctuations, and thermal exchanges. We investigate biological mechanisms for protein positioning and patterning within membranes…

DCPO: Dynamic Clipping Policy Optimization
Shihui Yang, Chengfeng Dou, Peidong Guo, Kai Lu, Qiang Ju, Fei Deng, Rihui Xin
https://arxiv.org/abs/2509.02333 https://

DCPO: Dynamic Clipping Policy Optimization
Reinforcement Learning from Verifiable Rewards (RLVR) has emerged as a promising framework for enhancing the reasoning capabilities of large language models. However, existing approaches such as GRPO often suffer from zero gradients. This problem arises primarily due to fixed clipping bounds for token-level probability ratios and the standardization of identical rewards, which can lead to ineffective gradient updates and underutilization of generated responses. In this work, we propose Dynamic …

Can SGD Handle Heavy-Tailed Noise?
Ilyas Fatkhullin, Florian H\"ubler, Guanghui Lan
https://arxiv.org/abs/2508.04860 https://arxiv.org/pdf/2508.04860

Can SGD Handle Heavy-Tailed Noise?
Stochastic Gradient Descent (SGD) is a cornerstone of large-scale optimization, yet its theoretical behavior under heavy-tailed noise -- common in modern machine learning and reinforcement learning -- remains poorly understood. In this work, we rigorously investigate whether vanilla SGD, devoid of any adaptive modifications, can provably succeed under such adverse stochastic conditions. Assuming only that stochastic gradients have bounded $p$-th moments for some $p \in (1, 2]$, we establish sha…

from my link log —
NaN-propagation: a novel method for sparsity detection in black-box computational functions.
https://arxiv.org/abs/2507.23186
saved 2025-08-14

NaN-Propagation: A Novel Method for Sparsity Detection in Black-Box Computational Functions
When numerically evaluating a function's gradient, sparsity detection can enable substantial computational speedups through Jacobian coloring and compression. However, sparsity detection techniques for black-box functions are limited, and existing finite-difference-based methods suffer from false negatives due to coincidental zero gradients. These false negatives can silently corrupt gradient calculations, leading to difficult-to-diagnose errors. We introduce NaN-propagation, which exploits the…

Particle-in-cell simulations of plasma wakefield formation in microwave waveguides
Jes\'us E. L\'opez, Eduardo A. Orozco
https://arxiv.org/abs/2506.17752

Particle-in-cell simulations of plasma wakefield formation in microwave waveguides
The acceleration of charged particles is fundamental not only for experimental studies in particle physics but also for applications in fields such as semiconductor manufacturing and medical therapies. However, conventional accelerators face limitations due to their large size, driven by low acceleration gradients. Plasma-based accelerators have emerged as a promising alternative, offering ultrahigh acceleration gradients, though their implementation is often limited by the need for high-intens…

Laser M\"ossbauer spectroscopy of ^{229}Th
Takahiro Hiraki, Takahiko Masuda, Sayuri Takatori, Fabian Schaden, Michael Bartokos, Kjeld Beeks, Yuta Fukunaga, Andreas Gr\"uneis, Ming Guan, Georgy Kazakov, Thomas LaGrange, Adrian Leitner, Ira Morawetz, Ryoichiro Ogake, Koichi Okai, Martin Pimon, Martin Pressler, Thomas Riebner, Noboru Sasao, Felix Schneider, Thorsten Schumm, Kotaro Shimizu, Luca Toscani de Col, Tomas Sikorsky, Akihiro Yoshimi, Koji Yoshimura

Laser Mössbauer spectroscopy of ^{229}Th
Mössbauer spectroscopy is widely used in biochemistry, geology, and solid-state physics to obtain structural information on materials. Here, we extend this technique into the optical range using a vacuum ultraviolet laser to probe the low-energy nuclear transitions of thorium-229, doped in calcium fluoride crystals. We discover four distinct doping sites for the thorium ions, determine the characteristic electric field gradients emerging in the interaction with the host crystal, and identify t…

Fast automated adjoints for spectral PDE solvers
Calum S. Skene, Keaton J. Burns
https://arxiv.org/abs/2506.14792 https://arxiv.org/p…

Fast automated adjoints for spectral PDE solvers
We present a general and automated approach for computing model gradients for PDE solvers built on sparse spectral methods, and implement this capability in the widely used open-source Dedalus framework. We apply reverse-mode automatic differentiation to symbolic graph representations of PDEs, efficiently constructing adjoint solvers that retain the speed and memory efficiency of this important class of modern numerical methods. This approach enables users to compute gradients and perform optim…

Boundary layer effects induced by the fluid in a chemotaxis-Navier-Stokes system
Qianqian Hou
https://arxiv.org/abs/2509.03028 https://arxiv.org/pdf/2509.0…

Boundary layer effects induced by the fluid in a chemotaxis-Navier-Stokes system
This paper is concerned with the boundary layer problem on a chemotaxis-Navier-Stokes system modelling the boundary layer formation of aerobic bacteria in fluids. Completing this system with Neumann boundary conditions on oxygen, we show that gradients of its second solution component in the half plane of $\mathbb{R}^2$ possess boundary layer effects as the oxygen diffusion rate goes to zero. However, neglecting the influence of the fluid, gradients of solutions to the chemotaxis-only subsystem…

Mass Transfer Through Vapor-Liquid Interfaces From Hydrodynamic Density Functional Theory
B. Bursik, F. Bender, R. Stierle, G. Bauer, J. Gross
https://arxiv.org/abs/2507.03017

Mass Transfer Through Vapor-Liquid Interfaces From Hydrodynamic Density Functional Theory
We assess the capabilities of hydrodynamic density functional theory (DFT) to predict mass transfer across vapor-liquid interfaces by studying the response of an initially equilibrated pure component vapor-liquid system to the localized insertion of a second component. Hydrodynamic DFT captures the effect of interfaces on the dynamics by modeling the chemical potential gradients of an inhomogeneous system based on classical DFT. Hydrodynamic DFT effectively connects molecular models with contin…

Temperature based radial metallicity gradients in nearby galaxies
K. Kreckel, R. J. Rickards Vaught, O. V. Egorov, J. E. M\'endez-Delgado, F. Belfiore, M. Brazzini, E. Egorova, E. Congiu, D. A. Dale, S. Dlamini, S. C. O. Glover, K. Grasha, R. S. Klessen, F. -H. Liang, H. -A. Pan, P. S\'anchez-Bl\'azquez, T. G Williams
https://a…

Temperature based radial metallicity gradients in nearby galaxies
Gas-phase abundances provide insights into the baryon cycle, with radial gradients and 2D metallicity distributions tracking how metals build up and redistribute within galaxy disks over cosmic time. We use a catalog of 22,958 HII regions across 19 nearby spiral galaxies to examine how precisely the radial abundance gradients can be traced using only the [NII]5755 electron temperature as a proxy for `direct method' metallicities. Using 534 direct detections of the temperature sensitive [NII]575…

Beyond Correctness: Harmonizing Process and Outcome Rewards through RL Training
Chenlu Ye, Zhou Yu, Ziji Zhang, Hao Chen, Narayanan Sadagopan, Jing Huang, Tong Zhang, Anurag Beniwal
https://arxiv.org/abs/2509.03403

Beyond Correctness: Harmonizing Process and Outcome Rewards through RL Training
Reinforcement learning with verifiable rewards (RLVR) has emerged to be a predominant paradigm for mathematical reasoning tasks, offering stable improvements in reasoning ability. However, Outcome Reward Models (ORMs) in RLVR are too coarse-grained to distinguish flawed reasoning within correct answers or valid reasoning within incorrect answers. This lack of granularity introduces noisy and misleading gradients significantly and hinders further progress in reasoning process quality. While Proc…

Equivariant machine learning of Electric Field Gradients -- Predicting the quadrupolar coupling constant in the MAPbI$_3$ phase transition
Bernhard Schmiedmayer (Jop), J. W. Wolffs (Jop), Gilles A. de Wijs, Arno P. M. Kentgens, Jonathan Lahnsteiner, Georg Kresse
https://arxiv.org/abs/2507.19435

Equivariant machine learning of Electric Field Gradients -- Predicting the quadrupolar coupling constant in the MAPbI$_3$ phase transition
We present a strategy combining machine learning and first-principles calculations to achieve highly accurate nuclear quadrupolar coupling constant predictions. Our approach employs two distinct machine-learning frameworks: a machine-learned force field to generate molecular dynamics trajectories and a second model for electric field gradients that preserves rotational and translational symmetries. By incorporating thermostat-driven molecular dynamics sampling, we enable the prediction of quadr…

Nambu Non-equilibrium Thermodynamics I: Foundation
So Katagiri, Yoshiki Matsuoka, Akio Sugamoto
https://arxiv.org/abs/2508.00207 https://arxiv.org/pdf/2508…

Nambu Non-equilibrium Thermodynamics I: Foundation
We propose a novel framework of non-equilibrium thermodynamics, termed Nambu Non-equilibrium Thermodynamics (NNET), which integrates reversible dynamics governed by the Nambu bracket with irreversible dynamics driven by entropy gradients. This framework is developed axiomatically and enables the description of far-from-equilibrium systems in which entropy can transiently decrease -- a regime difficult to handle within conventional approaches such as Onsager's linear response theory, Prigogine's…

Towards Performatively Stable Equilibria in Decision-Dependent Games for Arbitrary Data Distribution Maps
Guangzheng Zhong, Yang Liu, Jiming Liu
https://arxiv.org/abs/2509.02619

Towards Performatively Stable Equilibria in Decision-Dependent Games for Arbitrary Data Distribution Maps
In decision-dependent games, multiple players optimize their decisions under a data distribution that shifts with their joint actions, creating complex dynamics in applications like market pricing. A practical consequence of these dynamics is the \textit{performatively stable equilibrium}, where each player's strategy is a best response under the induced distribution. Prior work relies on $β$-smoothness, assuming Lipschitz continuity of loss function gradients with respect to the data distribu…

Local three-dimensional simulations of the convective overstability in protoplanetary discs
Robert J. Teed, Henrik N. Latter
https://arxiv.org/abs/2508.03557 https://

Local three-dimensional simulations of the convective overstability in protoplanetary discs
At certain radii protoplanetary discs may sustain a form of oscillatory convection (`convective overstability'; COS) due to localised adverse entropy gradients. The resulting hydrodynamical activity can produce coherent structures, such as zonal flows and vortices, that may concentrate solid material and aid their further coagulation. In this paper we extend previous axisymmetric runs by performing local three-dimensional simulations of the COS, using the code SNOOPY. As parameters are varied, …

Beyond Rate Coding: Surrogate Gradients Enable Spike Timing Learning in Spiking Neural Networks
Ziqiao Yu, Pengfei Sun, Dan F. M. Goodman
https://arxiv.org/abs/2507.16043 https:…

Beyond Rate Coding: Surrogate Gradients Enable Spike Timing Learning in Spiking Neural Networks
We investigate the extent to which Spiking Neural Networks (SNNs) trained with Surrogate Gradient Descent (Surrogate GD), with and without delay learning, can learn from precise spike timing beyond firing rates. We first design synthetic tasks isolating intra-neuron inter-spike intervals and cross-neuron synchrony under matched spike counts. On more complex spike-based speech recognition datasets (Spiking Heidelberg Digits (SHD) and Spiking Speech Commands (SSC), we construct variants where spi…

Replaced article(s) found for cond-mat.mes-hall. https://arxiv.org/list/cond-mat.mes-hall/new
[2/2]:
- Measuring high field gradients of cobalt nanomagnets in a spin-mechanical setup
Hahne, Pfau, \v{Z}aper, Stefan, Capelle, Ranfagni, Poggio, Schliesser

Cathodes and Shape Modification of Cavity for DESY Superconducting Photoinjector
J. Sekutowicz, D. Bazyl, E. Vogel, D. Reschke, A. Brinkmann, D. Klinke, D. Kostin, T. Ramm, A. Sulimov, H. Weise, M. Wiencek
https://arxiv.org/abs/2507.01658

Cathodes and Shape Modification of Cavity for DESY Superconducting Photoinjector
Four DESY prototypes of the L-band superconducting RF photoinjector cavity demonstrated on-axis peak gradients above 55MV/m during multiple vertical cryogenic tests. Two of these prototypes,16G09 and 16G10, achieved these gradients with both superconducting and normal-conducting metallic cathodes, fabricated from either high-purity niobium or lower-purity copper. The DESY photoinjector, under development for over two decades as a continuous-wave electron source for FELs, differs from other SRF …

Crosslisted article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[4/7]:
- On Task Vectors and Gradients
Zhou, Solombrino, Crisostomi, Bucarelli, D'Inverno, Silvestri, Rodol\`a

A unipolar head gradient for high-field MRI without encoding ambiguity
Markus Weiger, Johan Overweg, Franciszek Hennel, Emily Louise Baadsvik, Samuel Bianchi, Oskar Bj\"orkqvist, Roger Luechinger, Jens Metzger, Eric Michael, Thomas Schmid, Lauro Singenberger, Urs Sturzenegger, Erik Oskam, Gerrit Vissers, Jos Koonen, Wout Schuth, Jeroen Koeleman, Martino Borgo Klaas Paul Pruessmann

A unipolar head gradient for high-field MRI without encoding ambiguity
Purpose: MRI gradients with a conventional, bipolar design generally face a trade-off between performance, encoding ambiguity, and circumventing the latter by means of RF selectivity. This problem is particularly limiting in cutting-edge brain imaging performed at field strengths >= 7T and using high-performance head gradients. Methods: To address this issue, the present work proposes to fundamentally eliminate the encoding ambiguity in head gradients by using a unipolar z-gradient design that …

Single spin exact gradients for the optimization of complex pulses and pulse sequences
Stella Slad, Burkhard Luy
https://arxiv.org/abs/2507.13557 https://

Single spin exact gradients for the optimization of complex pulses and pulse sequences
The efficient computer optimization of magnetic resonance pulses and pulse sequences involves the calculation of a problem-adapted cost function as well as its gradients with respect to all controls applied. The gradients generally can be calculated as a finite difference approximation, as a GRAPE approximation, or as an exact function, e.g. by the use of the augmented matrix exponentiation, where the exact gradient should lead to best optimization convergence. However, calculation of exact gra…

A Melody in the Noise: Modeling Echoes of the Crab Nebula
Thierry Serafin Nadeau, Marten H. van Kerkwijk
https://arxiv.org/abs/2507.23214 https://arxiv.org…

A Melody in the Noise: Modeling Echoes of the Crab Nebula
Pulses from the Crab pulsar are often followed by ``echoes'', produced by radiation that was deflected by small filamentary structures in the Crab nebula and thus traveled via longer paths. We describe a simplified but detailed model that treats the filaments as cylinders of dense, neutral material with a thin ionized skin. In this picture, echoes are produced when the line of sight crosses the skin at glancing incidence, which naturally leads to the large electron column density gradients requ…

Can Classical Initialization Help Variational Quantum Circuits Escape the Barren Plateau?
Yifeng Peng, Xinyi Li, Zhemin Zhang, Samuel Yen-Chi Chen, Zhiding Liang, Ying Wang
https://arxiv.org/abs/2508.18497

Can Classical Initialization Help Variational Quantum Circuits Escape the Barren Plateau?
Variational quantum algorithms (VQAs) have emerged as a leading paradigm in near-term quantum computing, yet their performance can be hindered by the so-called barren plateau problem, where gradients vanish exponentially with system size or circuit depth. While most existing VQA research employs simple Gaussian or zero-initialization schemes, classical deep learning has long benefited from sophisticated weight initialization strategies such as Xavier, He, and orthogonal initialization to improv…

Diffusioosmotic flow in a soft microfluidic configuration induces fluid-structure instability
Nataly Maroundik, Dotan Ilssar, Evgeniy Boyko
https://arxiv.org/abs/2507.19065 http…

Diffusioosmotic flow in a soft microfluidic configuration induces fluid-structure instability
Diffusioosmotic flow arises in microfluidic configurations due to solute concentration gradients. In soft microfluidic channels, internal pressure gradients generated by diffusioosmotic flow to conserve mass result in elastic deformation of the channel walls, triggering fluid-structure interaction. In this work, we analyze the fluid-structure interaction between diffusioosmotic flow of an electrolyte solution and a deformable microfluidic channel. We provide insight into the physical behavior o…

Thermal-phototactic bioconvection in a forward scattering algal suspension
S. K. Rajput, M. K. Panda, A. Rathi
https://arxiv.org/abs/2506.23224 https://

Thermal-phototactic bioconvection in a forward scattering algal suspension
Bioconvection induced by phototaxis and thermal gradients in an anisotropic (forward) scattering algal suspension is investigated in this article. The suspension is illuminated by collimated irradiation from above and heated either from top or bottom. The linear theory is deployed on the steady state of the proposed bioconvective system and resulting eigen value problem is solved using fourth-order accurate finite-difference scheme based on Newton-Raphson-Kantorovich iteration. The results indi…

Local Differential Privacy for Distributed Stochastic Aggregative Optimization with Guaranteed Optimality
Ziqin Chen, Yongqiang Wang
https://arxiv.org/abs/2506.15106

Local Differential Privacy for Distributed Stochastic Aggregative Optimization with Guaranteed Optimality
Distributed aggregative optimization underpins many cooperative optimization and multi-agent control systems, where each agent's objective function depends both on its local optimization variable and an aggregate of all agents' optimization variables. Existing distributed aggregative optimization approaches typically require access to accurate gradients of the objective functions, which, however, are often hard to obtain in real-world applications. For example, in machine learning, gradients ar…

O1NumHess: a fast and accurate seminumerical Hessian algorithm using only O(1) gradients
Bo Wang, Shaohang Luo, Zikuan Wang, Wenjian Liu
https://arxiv.org/abs/2508.07544 https:/…

O1NumHess: a fast and accurate seminumerical Hessian algorithm using only O(1) gradients
In this work, we describe a new algorithm, O1NumHess, to calculate the Hessian of a molecular system by finite differentiation of gradients calculated at displaced geometries. Different from the conventional seminumerical Hessian algorithm, which requires gradients at $O(N_{\mathrm{atom}})$ displaced geometries (where $N_{\mathrm{atom}}$ is the number of atoms), the present approach only requires $O(1)$ gradients. Key to the reduction of the number of gradients is the exploitation of the off-di…

Active sorting to boundaries in active nematic -- passive isotropic fluid mixtures
Saraswat Bhattacharyya, Julia M. Yeomans
https://arxiv.org/abs/2509.01523 https://

Active sorting to boundaries in active nematic -- passive isotropic fluid mixtures
We use a two-fluid model to study a confined mixture of an active nematic fluid and a passive isotropic fluid. We find that an extensile active fluid preferentially accumulates at a boundary if the anchoring is planar, whereas its boundary concentration decreases for homeotropic anchoring. These tendencies are reversed if the active fluid is contractile. We argue that the sorting results from gradients in the nematic order, and show that the behaviour can be driven by either imposed boundary an…

GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
Lakshya A Agrawal, Shangyin Tan, Dilara Soylu, Noah Ziems, Rishi Khare, Krista Opsahl-Ong, Arnav Singhvi, Herumb Shandilya, Michael J Ryan, Meng Jiang, Christopher Potts, Koushik Sen, Alexandros G. Dimakis, Ion Stoica, Dan Klein, Matei Zaharia, Omar Khattab
https://arx…

GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
Large language models (LLMs) are increasingly adapted to downstream tasks via reinforcement learning (RL) methods like Group Relative Policy Optimization (GRPO), which often require thousands of rollouts to learn new tasks. We argue that the interpretable nature of language can often provide a much richer learning medium for LLMs, compared with policy gradients derived from sparse, scalar rewards. To test this, we introduce GEPA (Genetic-Pareto), a prompt optimizer that thoroughly incorporates …

Local boundary conditions in nonlocal hyperelasticity via heterogeneous horizons
Carolin Kreisbeck, Hidde Sch\"onberger
https://arxiv.org/abs/2509.03468 https://

Local boundary conditions in nonlocal hyperelasticity via heterogeneous horizons
In this paper, we consider a class of variational problems with integral functionals involving nonlocal gradients. These models have been recently proposed as refinements of classical hyperelasticity, aiming for an effective framework to capture also discontinuous and singular material effects. Specific to our set-up is a space-dependent interaction range that vanishes at the boundary of the reference domain. This ensures that the nonlocal operator depends only on values within the domain and l…

Crosslisted article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[4/7]:
- On Task Vectors and Gradients
Zhou, Solombrino, Crisostomi, Bucarelli, D'Inverno, Silvestri, Rodol\`a

3D Simulations of Convective Entrainment in Gas Giants: Rotation and Decreasing Luminosity as Barriers to Mixing
Shu Zhang, J. Rafael Fuentes, Andrew Cumming
https://arxiv.org/abs/2509.00364

3D Simulations of Convective Entrainment in Gas Giants: Rotation and Decreasing Luminosity as Barriers to Mixing
Observations from Juno and Cassini suggest that Jupiter and Saturn may possess fuzzy cores -- central regions where the abundance of heavy elements varies smoothly with depth. Such gradients pose a longstanding puzzle for models of planetary evolution and formation, which predict that vigorous convection would homogenize the interior of gas giants within the first $\sim 10^6$--$10^8~\mathrm{years}$ of cooling. Previous 3D simulations and analytic predictions for the propagation of a convection …

Flexoelectric Polarization Enhancement in Paraelectric $\mathrm{BaHfO_3}$ via Strain Gradient Engineering
Timo Piecuch, Jeffrey A. Brock, Xiaochun Huang, Arnold M. M\"uller, Christof Vockenhuber, Christof W. Schneider, Thomas Lippert, Nick A. Shepelin
https://arxiv.org/abs/2506.19690

Flexoelectric Polarization Enhancement in Paraelectric $\mathrm{BaHfO_3}$ via Strain Gradient Engineering
Flexoelectricity - polarization induced by strain gradients - offers a route to polar functionality in centrosymmetric dielectrics, where traditional piezoelectric effects are absent. This study investigates the flexoelectric effect in epitaxial $\mathrm{BaHfO_3}$ (BHO) thin films, a centrosymmetric and paraelectric perovskite. While a large lattice mismatch induces defect-driven relaxation, a coherently grown BHO film undergoes elastic relaxation, forming intrinsic strain gradients exceeding $…

Signature of Pressure Gradient History on Wall Shear Stress in Turbulent Boundary Layers
Marco Mattei, Theresa Saxton-Fox
https://arxiv.org/abs/2507.01244 …

Signature of Pressure Gradient History on Wall Shear Stress in Turbulent Boundary Layers
Our experiments evaluate the presence of history effects on wall shear stress in turbulent boundary layers (TBLs) subjected to five sequences of favorable and adverse pressure gradients (FAPGs) of increasing strength. Steady but spatially varying FAPGs are imposed on a flat plate using a deformable convex false ceiling. An array of three one-dimensional capacitive probes, positioned in the adverse pressure gradient (APG) region downstream of the favorable pressure gradient (FPG), enables direct…

Controlling Skyrmion Lattice Orientation with Local Magnetic Field Gradients
Duc Minh Tran, Edoardo Mangini, Elizabeth M. Jefremovas, Fabian Kammerbauer, Dennis Meier, Robert Fr\"omter, Mathias Kl\"aui
https://arxiv.org/abs/2508.14771

Controlling Skyrmion Lattice Orientation with Local Magnetic Field Gradients
Precise control over the formation and arrangement of magnetic skyrmion lattices is essential for understanding their emergent behavior and advancing their integration into spintronic and magnonic devices. We report on a simple and minimally invasive technique to nucleate and manipulate skyrmion lattices in soft magnetic CoFeB using single-pass magnetic force microscopy (MFM). By tuning the scan-line spacing to match the intrinsic stripe domain periodicity, the stray field gradient from the MFM…

Metal-THINGS: gas metallicity gradients in nearby galaxies
Guillermo Val\'e, Maritza A. Lara-L\'opez, Mabel Valerdi, Igor Zinchenko, Shane P. O'Sullivan, Leonid S. Pilyugin, Jordi Cepa, Viviana Casasola, Mar\'ia E. De Rossi, Sami Dib, Jacopo Fritz, Jes\'us Gallego, Luis E. Gardu\~no, Omar L\'opez-Cruz, Vidhi Tailor, Javier Zaragoza-Cardiel

Metal-THINGS: gas metallicity gradients in nearby galaxies
This paper explores the gas metallicity gradients in a sample of 25 nearby galaxies using new Integral Field Spectroscopy observations from the Metal-THINGS survey. We derive and study the resolved diffuse ionised gas content, Baldwin, Phillips and Terlevich diagrams and gas metallicities for our entire sample, at spatial resolutions of 40-300 pc. Gas metallicity gradients are studied as a function of the galaxy's stellar mass, H I gas fraction, diffuse ionised gas content, and using different …

SAGE: A Set-based Adaptive Gradient Estimator
Lorenzo Sabug Jr., Fredy Ruiz, Lorenzo Fagiano
https://arxiv.org/abs/2508.19400 https://arxiv.org/pdf/2508.19…

SAGE: A Set-based Adaptive Gradient Estimator
A new paradigm to estimate the gradient of a black-box scalar function is introduced, considering it as a member of a set of admissible gradients that are computed using existing function samples. Results on gradient estimate accuracy, derived from a multivariate Taylor series analysis, are used to express the set of admissible gradients through linear inequalities. An approach to refine this gradient estimate set to a desired precision is proposed as well, using an adaptive sampling approach. …

Strategic Incentivization for Locally Differentially Private Federated Learning
Yashwant Krishna Pagoti, Arunesh Sinha, Shamik Sural
https://arxiv.org/abs/2508.07138 https://

Strategic Incentivization for Locally Differentially Private Federated Learning
In Federated Learning (FL), multiple clients jointly train a machine learning model by sharing gradient information, instead of raw data, with a server over multiple rounds. To address the possibility of information leakage in spite of sharing only the gradients, Local Differential Privacy (LDP) is often used. In LDP, clients add a selective amount of noise to the gradients before sending the same to the server. Although such noise addition protects the privacy of clients, it leads to a degrada…

Confocal Raman microscopy inside sessile multicomponent droplets
Alexander Erb, Johanna Steinmann, Youngeun Lee, Robert W. Stark
https://arxiv.org/abs/2507.00208

Confocal Raman microscopy inside sessile multicomponent droplets
Evaporating multicomponent droplets are ubiquitous and appear in a variety of everyday situations and technological applications, including coating, 3D printing, and energy conversion processes. During evaporation, concentration gradients are typically induced, resulting in flows within the droplets. Many of the mechanisms underlying multicomponent droplet evaporation are not fully understood. However, most methods utilize markers that can be surface active and thus affect droplet dynamics. Thu…

Trust-region Filter Algorithms utilising Hessian Information for Grey-Box Optimisation
Gul Hameed, Tao Chen, Antonio del Rio Chanona, Lorenz T. Biegler, Michael Short
https://arxiv.org/abs/2509.01651

Trust-region Filter Algorithms utilising Hessian Information for Grey-Box Optimisation
Optimising industrial processes is essential for enhancing profitability, efficiency, and sustainability. Conventional optimisation relies on equation-oriented (EO) glass-box models with fully specified equations and accessible derivatives. However, many real-world systems contain black-box components (such as high-fidelity computational fluid dynamics simulators or proprietary unit models) that lack explicit formulations or analytic gradients. This limits the applicability of derivative-based …

DNT: a Deeply Normalized Transformer that can be trained by Momentum SGD
Xianbiao Qi, Marco Chen, Wenjie Xiao, Jiaquan Ye, Yelin He, Chun-Guang Li, Zhouchen Lin
https://arxiv.org/abs/2507.17501

DNT: a Deeply Normalized Transformer that can be trained by Momentum SGD
Transformers have become the de facto backbone of modern deep learning, yet their training typically demands an advanced optimizer with adaptive learning rate like AdamW, rather than a momentum SGDW (mSGDW). Previous works show that it is mainly due to a heavy-tailed distribution of the gradients. In this paper, we introduce a Deeply Normalized Transformer (DNT), which is meticulously engineered to overcome this limitation enabling seamless training with vanilla mSGDW while yielding comparable …

Extensional rheology of dilute suspensions of spheres in polymeric liquids
Arjun Sharma, Donald L. Koch
https://arxiv.org/abs/2507.23114 https://arxiv.org/…

Extensional rheology of dilute suspensions of spheres in polymeric liquids
The extensional rheology of dilute suspensions of spheres in viscoelastic or polymeric liquids is studied computationally. At low polymer concentration (c) and Deborah number (De), a wake of highly stretched polymers forms downstream of the particles due to larger local velocity gradients than the imposed flow, indicated by a positive deviation in local De. This increases the suspension's extensional viscosity with time and De for De less than 0.5. When De exceeds 0.5 (the coil-stretch transiti…

Proving the Limited Scalability of Centralized Distributed Optimization via a New Lower Bound Construction
Alexander Tyurin
https://arxiv.org/abs/2506.23836

Proving the Limited Scalability of Centralized Distributed Optimization via a New Lower Bound Construction
We consider centralized distributed optimization in the classical federated learning setup, where $n$ workers jointly find an $\varepsilon$-stationary point of an $L$-smooth, $d$-dimensional nonconvex function $f$, having access only to unbiased stochastic gradients with variance $σ^2$. Each worker requires at most $h$ seconds to compute a stochastic gradient, and the communication times from the server to the workers and from the workers to the server are $τ_{s}$ and $τ_{w}$ seconds per coo…

ZORMS-LfD: Learning from Demonstrations with Zeroth-Order Random Matrix Search
Olivia Dry, Timothy L. Molloy, Wanxin Jin, Iman Shames
https://arxiv.org/abs/2507.17096

ZORMS-LfD: Learning from Demonstrations with Zeroth-Order Random Matrix Search
We propose Zeroth-Order Random Matrix Search for Learning from Demonstrations (ZORMS-LfD). ZORMS-LfD enables the costs, constraints, and dynamics of constrained optimal control problems, in both continuous and discrete time, to be learned from expert demonstrations without requiring smoothness of the learning-loss landscape. In contrast, existing state-of-the-art first-order methods require the existence and computation of gradients of the costs, constraints, dynamics, and learning loss with re…

Quantifying non-equilibrium pressure-gradient turbulent boundary layers through a symmetry-based framework
Wei-Tao Bi, Ke-Xin Zheng, Jun Chen, Zhen-Su She
https://arxiv.org/abs/2508.21447

Quantifying non-equilibrium pressure-gradient turbulent boundary layers through a symmetry-based framework
This study establishes a symmetry-based framework to quantify non-equilibrium processes in complex pressure gradient (PG) turbulent boundary layers (TBLs), using a Lie-group-informed dilation-symmetry-breaking formalism. We derive a universal multilayer defect scaling law for the evolution of total shear stress (TSS). The law shows that gradually varying adverse pressure gradients (APGs) break the dilation symmetry in the two-layer defect scaling of equilibrium TSS, leading to three-layer TSS s…

A Structured Proximal Stochastic Variance Reduced Zeroth-order Algorithm
Marco Rando, Cheik Traor\'e, Cesare Molinari, Lorenzo Rosasco, Silvia Villa
https://arxiv.org/abs/2506.23758

A Structured Proximal Stochastic Variance Reduced Zeroth-order Algorithm
Minimizing finite sums of functions is a central problem in optimization, arising in numerous practical applications. Such problems are commonly addressed using first-order optimization methods. However, these procedures cannot be used in settings where gradient information is unavailable. Finite-difference methods provide an alternative by approximating gradients through function evaluations along a set of directions. For finite-sum minimization problems, it was shown that incorporating varian…

Viscosity variation in fluid flows across scale
Arjun Sharma, Ritabrata Thakur, Sharath Jose, Rama Govindarajan
https://arxiv.org/abs/2507.23532 https://ar…

Viscosity variation in fluid flows across scale
A wide range of natural and engineered fluid flows exhibit spatial or temporal viscosity variations, spanning scales from microbial locomotion to planetary mantle convection. These variations introduce qualitatively new physical mechanisms absent in constant-viscosity flows. This review surveys such phenomena across scales. In low Reynolds number (Stokes) flows, viscosity gradients couple translation and rotation, enabling novel particle responses to uniform forcing-- mechanisms that microorgan…

Set Invariance with Probability One for Controlled Diffusion: Score-based Approach
Wenqing Wang, Alexis M. H. Teter, Murat Arcak, Abhishek Halder
https://arxiv.org/abs/2507.22385

Set Invariance with Probability One for Controlled Diffusion: Score-based Approach
Given a controlled diffusion and a connected, bounded, Lipschitz set, when is it possible to guarantee controlled set invariance with probability one? In this work, we answer this question by deriving the necessary and sufficient conditions for the same in terms of gradients of certain log-likelihoods -- a.k.a. score vector fields -- for two cases: given finite time horizon and infinite time horizon. The deduced conditions comprise a score-based test that provably certifies or falsifies the exi…