arXiv AI Agent for Weinan Wang

Your research scout briefing

agent triaged 61 papers; briefed 24

Physics-guided correction for operator learning under model misspecification

read now
relevance ★★★★★
math.NA  ·  2606.03469

Lei Ma, Nicolas Boullé, Yu-Sen Yang, Hao Wu, Ling Guo

Scout score 20.6   title matches operator learning, learning, operator / watch term operator learning, drift / category weight math.NA +1.5

Physics-informed operator learning provides an efficient framework for approximating solution operators of partial differential equations by combining observational data with governing physical laws. However, most existing methods implicitly assume that the prescribed governing equation is accurate.

Connection to your workThe local scout matched this paper to your profile through operator learning, learning, operator, drift, operators.

Next action: Read the introduction and main theorem statements, focusing on operator learning, learning, operator.

When Both Layers Learn: Training Dynamics of Representing Linear Models via ReLU Networks

read now
relevance ★★★★★
cs.LG · math.OC · math.ST · stat.ML  ·  2606.04476

Berk Tinaz, Changzhi Xie, Mahdi Soltanolkotabi

Scout score 19.3   title matches training, networks, when / watch term inverse problems / category weight cs.LG +0.8, math.OC +2.0, math.ST +1.5

In this paper, we study the gradient descent dynamics for jointly training both layers of a one-hidden-layer ReLU network to fit a linear target function. Concretely, we consider a realizable setting where inputs are drawn i.i.d.

Connection to your workThe local scout matched this paper to your profile through training, networks, when, inverse problems, inverse.

Next action: Read the introduction and main theorem statements, focusing on training, networks, when.

Let There Be Light: Reflection, Refraction and Scattering for Neural Operators

read now
relevance ★★★★★
cs.LG · math.NA  ·  2606.03262

Keke Wu, Yixuan Zhang, Jingrun Chen

Scout score 18.4   title matches neural operator, operators, neural / watch term neural operator / category weight cs.LG +0.8, math.NA +1.5

Neural operators learn mappings between infinite-dimensional function spaces and provide a data-driven surrogate modeling paradigm for parametric partial differential equations (PDEs). Existing architectures typically obtain expressivity by parameterizing integral kernels in prescribed transform domains or by applying attention-like interactions over discretized spatial points.

Connection to your workThe local scout matched this paper to your profile through neural operator, operators, neural, learning, operator.

Next action: Read the introduction and main theorem statements, focusing on neural operator, operators, neural.

Bayesian learning for the stochastic shortest path problem

read now
relevance ★★★★★
stat.ML · cs.LG · math.ST · stat.CO  ·  2606.04845

Chon Wai Ho, Sumeetpal S. Singh, Jiaqi Guo

Scout score 15.0   title matches learning, bayesian, stochastic / category weight stat.ML +1.0, cs.LG +0.8, math.ST +1.5

Sequential decision-making problems are often modelled as a Markov decision process (MDP). We focus on the stochastic shortest path (SSP) problem, which is an infinite-horizon undiscounted MDP with absorbing terminal states.

Connection to your workThe local scout matched this paper to your profile through learning, bayesian, stochastic, optimal, uncertainty.

Next action: Read the introduction and main theorem statements, focusing on learning, bayesian, stochastic.

Boundary-compatible interacting approximations of quasilinear PDEs on bounded domains

read now
relevance ★★★★★
math.NA  ·  2606.04049

Thierry Paul, Emmanuel Trélat

Scout score 14.4   title matches pdes, boundary / watch term drift / category weight math.NA +1.5

We develop a general operator-theoretic route that turns Kato-type quasilinear evolution systems on a Banach scale $(Z,X)$ into finite-dimensional interacting approximations. The construction proceeds in two steps.

Connection to your workThe local scout matched this paper to your profile through pdes, boundary, drift, operators, operator.

Next action: Read the introduction and main theorem statements, focusing on pdes, boundary, drift.

Nonlocal Mean Field Schrödinger Bridge with Learned Interactions

read now
relevance ★★★★★
math.OC · cs.LG · math.NA  ·  2606.04265

Daisuke Inoue, Mathieu Laurière, Dante Kalise

Scout score 14.2   title matches mean, field / category weight math.OC +2.0, cs.LG +0.8, math.NA +1.5

The Schrödinger Bridge Problem constructs a stochastic process that connects an initial distribution to a terminal distribution with minimum energy. This work considers its mean-field extension, the Mean-Field Schrödinger Bridge, for interacting particle systems.

Connection to your workThe local scout matched this paper to your profile through mean, field, stochastic, training, neural.

Next action: Read the introduction and main theorem statements, focusing on mean, field, stochastic.

Neural Networks Provably Learn Spectral Representations for Group Composition

read now
relevance ★★★★
cs.LG · math.OC · math.RT · math.ST  ·  2606.02993

Jianliang He, Leda Wang, Fengzhuo Zhang, Siyu Chen, Zhuoran Yang

Scout score 12.4   title matches neural, networks / category weight cs.LG +0.8, math.OC +2.0, math.ST +1.5

Understanding how structured internal structure emerges during neural network training is central to the study of deep learning. We investigate this phenomenon through the group composition task, where a two-layer neural network is trained to predict $g_1 \star g_2$ for elements of a finite group $G$.

Connection to your workThe local scout matched this paper to your profile through neural, networks, learning, training.

Next action: Read the introduction and main theorem statements, focusing on neural, networks, learning.

Deep Embedded Multiplicative DMD for Algebra-Preserving Koopman Learning

read now
relevance ★★★★
cs.LG · math.DS · math.NA · math.OC  ·  2606.05131

Kelan Gray, Finlay Brown, Nicolas Boullé, Matthew J. Colbrook

Scout score 11.7   title matches learning / category weight cs.LG +0.8, math.NA +1.5, math.OC +2.0

Koopman theory turns nonlinear dynamics into a linear spectral problem. In computation, however, everything depends on a hard finite-dimensional choice: the observables must be expressive, nearly invariant under the dynamics, and, ideally, compatible with composition.

Connection to your workThe local scout matched this paper to your profile through learning, operator, training, fluid, compact.

Next action: Read the introduction and main theorem statements, focusing on learning, operator, training.

A Fast Screening Approach for High-dimensional Outcomes and High-dimensional Predictors

read now
relevance ★★★★
stat.ME · cs.LG · math.ST · stat.ML  ·  2606.03018

Hongju Park, Zhenyao Ye, Shuo Chen

Scout score 11.0   title matches high / category weight stat.ME +1.0, cs.LG +0.8, math.ST +1.5

Modeling interactions among multimodal, high-dimensional data is intrinsically challenging due to ultra-high dimensionality and complex dependence structure with high level noise. Screening methods are effective for reducing dimensionality, but most existing approaches shrink only the predictor space while retaining all outcomes.

Connection to your workThe local scout matched this paper to your profile through high, applied, modeling, response, dependence.

Next action: Read the introduction and main theorem statements, focusing on high, applied, modeling.

Structure-Preserving Quantum Method of Lines for Evolutionary PDEs with Mixed Boundary Conditions

read now
relevance ★★★★
quant-ph · math.NA  ·  2606.03407

Yixuan Liang, Jin-Peng Liu

Scout score 10.0   title matches pdes, boundary / category weight math.NA +1.5

We give detailed analysis and circuit design of structure-preserving quantum algorithms for second-order linear evolutionary PDEs, including parabolic equations and hyperbolic equations with mixed Dirichlet, Neumann, and periodic boundary conditions and source terms. While prior quantum algorithms usually neglect the stability problem from the PDE-to-ODE reduction, our method-of-lines approach investigates the boundary lifting via Coons interpolation and boundary-aware discretization, so that the resulting...

Connection to your workThe local scout matched this paper to your profile through pdes, boundary, optimal, parabolic, systems.

Next action: Read the introduction and main theorem statements, focusing on pdes, boundary, optimal.

Structured drift design for denoising diffusion models

read now
relevance ★★★★
math.ST  ·  2606.03360

Mahsa Taheri

Scout score 9.8   title matches drift / watch term drift / category weight math.ST +1.5

Diffusion-based generative models have achieved remarkable success in high-dimensional data generation; however, they fundamentally rely on isotropic diffusion processes that destroy meaningful geometric structures in the forward process. For complex, multimodal, and highly correlated distributions such as biologically constrained genetic data, isotropic noise merges distinct modes and distorts intrinsic dependencies.

Connection to your workThe local scout matched this paper to your profile through drift, high, data, processes.

Next action: Read the introduction and main theorem statements, focusing on drift, high, data.

High-Order Summation-By-Parts Schemes for First-Order Hyperbolic Systems in Curvilinear Coordinates with Singularities

read now
relevance ★★★★
gr-qc · math.NA  ·  2606.05155

Stamatis Vretinaris, Erik Schnetter

Scout score 9.3   title matches systems, high / category weight math.NA +1.5

Formulating stable numerical methods for hyperbolic systems in curvilinear coordinate with singularities, e.g. spherical coordinates, is complicated by the presence of these singularities.

Connection to your workThe local scout matched this paper to your profile through systems, high, operators, scalar.

Next action: Read the introduction and main theorem statements, focusing on systems, high, operators.

Robust Prediction Variance Estimation for Gaussian Process Regression Under Covariance Smoothness Misspecification

read now
relevance ★★★★
stat.ME · math.ST  ·  2606.04322

Roberto Rivera

Scout score 9.2   title matches prediction / category weight stat.ME +1.0, math.ST +1.5

Best Linear Unbiased Prediction (BLUP) has been a dominant approach in Generalized Linear Mixed Models, spatial models, and Gaussian Process Regression (GPR). In addition to their optimal properties, BLUP procedures quantify prediction uncertainty.

Connection to your workThe local scout matched this paper to your profile through prediction, optimal, uncertainty, mean, probability.

Next action: Read the introduction and main theorem statements, focusing on prediction, optimal, uncertainty.

Well-posedness and finite element approximation of the electrostatic shear Alfvén wave equations

read now
relevance ★★★★
math.NA · math.AP  ·  2606.04644

Annalisa Buffa, Timon Miehling, Marco Picasso, Markus Renoldner

Scout score 9.1   abstract matches drift, sobolev, field / watch term drift / category weight math.NA +1.5, math.AP +2.5

The aim of this paper is to study the well-posedness and finite element approximation of the electrostatic shear Alfvén wave equations, a coupled system of two partial differential equations arising in plasma physics as a simplified sub-model of the drift-reduced Braginskii equations. To this end, anisotropic Sobolev spaces depending on the normalized magnetic field $\b$ are introduced, together with a Poincaré-type inequality along the integral curves of $\b$, which holds under a geometric directedness condition...

Connection to your workThe local scout matched this paper to your profile through drift, sobolev, field.

Next action: Read the introduction and main theorem statements, focusing on drift, sobolev, field.

Critical evaluation of PINN for FWD inverse analysis and differentiable FEM as an alternative

read now
relevance ★★★★
cs.CE · cs.LG · math.NA  ·  2606.03210

Yongjin Choi, Hyeonbin Moon, Seunghwa Ryu

Scout score 9.0   title matches inverse / category weight cs.LG +0.8, math.NA +1.5

Automatic-differentiation-based inverse analysis methods, including physics-informed neural networks (PINNs) and differentiable programming, have recently shown great promise due to their ability to compute accurate gradients and convergence efficiency. However, their applicability to falling weight deflectometer (FWD) backcalculation remains unexplored.

Connection to your workThe local scout matched this paper to your profile through inverse, neural, systems, weighting, networks.

Next action: Read the introduction and main theorem statements, focusing on inverse, neural, systems.

Bregman divergences and error control via convex duality

read now
relevance ★★★☆☆
math.NA · math.AP  ·  2606.05088

P. A. Gazca-Orozco

Scout score 8.6   title matches control / category weight math.NA +1.5, math.AP +2.5

Convex duality relations are a useful tool for deriving error estimates for challenging nonlinear and non-smooth variational problems. Applied at the continuous level they can deliver nonlinear analogues of the Prager-Synge a posteriori error identity, while at the discrete level they allow the derivation of minimal regularity a priori estimates.

Connection to your workThe local scout matched this paper to your profile through control, applied, regularity.

Next action: Read the introduction and main theorem statements, focusing on control, applied, regularity.

A novel class of high-order uniformly accurate exponential integrators with local linear extension for the charged-particle dynamics under strong magnetic field

read now
relevance ★★★☆☆
math.NA  ·  2606.04478

Lina Wang, Bin Wang, Beibei Zhu

Scout score 8.6   title matches field, high / category weight math.NA +1.5

In this paper, we develop a novel class of high-order uniformly accurate exponential integrators for charged-particle dynamics under a strong magnetic field. The small parameter $0<\varepsilon\ll 1$ induces rapid temporal oscillations, rendering traditional numerical methods prohibitively expensive due to severe step-size restrictions.

Connection to your workThe local scout matched this paper to your profile through field, high, applied.

Next action: Read the introduction and main theorem statements, focusing on field, high, applied.

A decoupled energy-stable mixed finite element method for Poisson-Nernst-Planck-Navier-Stokes equations

read now
relevance ★★★☆☆
math.NA  ·  2606.04941

Yanmei Peng, Ying Yang, Yi Zhang

Scout score 8.4   title matches navier-stokes / watch term navier-stokes / category weight math.NA +1.5

We propose a novel linearized mixed finite element method for the Poisson-Nernst-Planck-Navier-Stokes (PNPNS) system. Specifically, the method combines a staggered time discretization that eliminates the need for expensive nonlinear solvers by carefully treating nonlinear terms in a time-staggered manner, with a mimetic spatial discretization that preserves the exact structure of the discrete de Rham complex.

Connection to your workThe local scout matched this paper to your profile through navier-stokes, data.

Next action: Read the introduction and main theorem statements, focusing on navier-stokes, data.

Decoupled iterative schemes for solving stationary MHD problems

read now
relevance ★★★☆☆
math.NA  ·  2606.04467

Aziz Takhirov, Driss Yakoubi

Scout score 8.4   title matches mhd / watch term mhd / category weight math.NA +1.5

We develop a novel iterative approach for solving the incompressible magnetohydrodynamics problem. The main idea is to split the velocity-momentum and magnetic induction equations with respect to the diffusive terms.

Connection to your workThe local scout matched this paper to your profile through mhd, systems, incompressible.

Next action: Read the introduction and main theorem statements, focusing on mhd, systems, incompressible.

Embedded Trefftz DG method for reaction-diffusion problems on anisotropic meshes

read now
relevance ★★★☆☆
math.NA  ·  2606.03845

Sergio Gómez, Chiara Perinati, Paul Stocker, Igor Voulis

Scout score 8.4   title matches reaction-diffusion / watch term reaction-diffusion / category weight math.NA +1.5

We present and analyze an embedded Trefftz discontinuous Galerkin method for reaction-diffusion problems on anisotropic meshes. The method is constructed by imposing a relaxed local Trefftz condition via an embedding into a tensor-product DG space, yielding a reduced global system while preserving the approximation properties of the underlying high-order discretization.

Connection to your workThe local scout matched this paper to your profile through reaction-diffusion, high.

Next action: Read the introduction and main theorem statements, focusing on reaction-diffusion, high.

Linear Convergence of Parareal Algorithm for Semilinear Parabolic Equations

read now
relevance ★★★☆☆
math.NA  ·  2606.03516

Guanglian Li, Qingle Lin, Shu-lin Wu, Zhi Zhou

Scout score 8.2   title matches parabolic / category weight math.NA +1.5

Long-time simulations of evolution equations present substantial computational challenges due to the inherently sequential nature of conventional time-stepping schemes. The parareal method, a leading parallel-in-time (PinT) algorithm, offers a promising approach to overcome the challenge by introducing concurrency in the time domain.

Connection to your workThe local scout matched this paper to your profile through parabolic, optimal, behavior, regularity, data.

Next action: Read the introduction and main theorem statements, focusing on parabolic, optimal, behavior.

An Energy-Stable Implicit Convex-Splitting BDF2 Scheme for the Cahn-Hilliard-Navier-Stokes Equations

skim
relevance ★★★☆☆
math.NA  ·  2606.04204

Xuelong Gu, Qi Wang

Scout score 7.7   title matches navier-stokes / watch term navier-stokes / category weight math.NA +1.5

We develop an energy-stable implicit convex-splitting BDF2 discretization (CS-BDF2) of the Cahn--Hilliard--Navier--Stokes equations. For the Cahn--Hilliard equation, BDF2 analyses can establish energy stability by testing the phase equation in the (H^{-1}) metric.

Connection to your workThe local scout matched this paper to your profile through navier-stokes, metric.

Next action: Skim the abstract, introduction, and conclusion for navier-stokes, metric.

On multi-fidelity methods for a tumor growth model with uncertainties

skim
relevance ★★★☆☆
math.NA  ·  2606.03607

Huimin Yu, Liu Liu, Yu Feng, Te Qi

Scout score 7.3   abstract matches uncertainty quantification, uncertainty, high / watch term uncertainty quantification / category weight math.NA +1.5

We develop a hierarchical multi-fidelity (MF) framework for efficient uncertainty quantification of porous-medium equation (PME) tumor growth models with moving free boundaries. The proposed approach combines coarse-grid PME solvers, level-set approximations of the Hele--Shaw limit, and fine-grid asymptotic-preserving PME discretizations, thereby integrating both discretization-based and asymptotic-model-based fidelity reduction.

Connection to your workThe local scout matched this paper to your profile through uncertainty quantification, uncertainty, high, quantification.

Next action: Skim the abstract, introduction, and conclusion for uncertainty quantification, uncertainty, high.

Fitting scattered data with optional monotonicity constraints on GPU: LipFit package

skim
relevance ★★★☆☆
math.NA · cs.LG · cs.MS  ·  2606.04670

Gleb Beliakov

Scout score 6.9   title matches data / category weight math.NA +1.5, cs.LG +0.8

This paper presents a method of multivariate scattered data interpolation and approximation that produces optimal Lipschitz-continuous approximation, subject to the desired monotonicity constraints. This method relies on tight upper and lower approximations to the data, and is similar in its spirit to the nearest-neighbour approximation but does not suffer from discontinuities.

Connection to your workThe local scout matched this paper to your profile through data, optimal, training.

Next action: Skim the abstract, introduction, and conclusion for data, optimal, training.

← archive