arXiv research

The comparison principle for scalar second order parabolic PDEs on functions $u(t,x)$ admits a topological interpretation: pairs of solutions, $u^1(t,\cdot)$ and $u^2(t,\cdot)$, evolve so as to not increase the intersection number of their graphs. We generalize to the case of multiple solutions $\{u^α(t,\cdot)\}_{α=1}^…

Paper uses deep Ritz method for solving stationary Schrödinger equation, proving convergence and feature emergence.

Study geometric flow on curves with positive torsion, finding stationary solutions and their stability.

Center manifold analysis can be used in order to investigate the stability of the stationary solutions of various PDEs. This can be done by considering the PDE as an ODE between certain Banach spaces and linearising about the stationary solution. Here we investigate the volume preserving mean curvature flow using such …

In this paper we study the reductions of evolutionary PDEs on the manifold of the stationary points of time--dependent symmetries. In particular we describe how that the finite dimensional Hamiltonian structure of the reduced system is obtained from the Hamiltonian structure of the initial PDE and we construct the time…

Develops a mean-field theory for multi-head self-attention under cross-entropy training.

Develops a PDE approach to constructing nontrivial anisotropic surfaces.

The class of non-rigid registration methods proposed in the framework of PDE-constrained Large Deformation Diffeomorphic Metric Mapping is a particularly interesting family of physically meaningful diffeomorphic registration methods. PDE-constrained LDDMM methods are formulated as constrained variational problems, wher…

A new method uses multifidelity Gaussian process regression to solve nonlinear PDEs.

Internal Lagrangians derived from variational principles.

New algorithm optimizes nonlinear SDEs online with convergence guarantees.

Turnpike property applies to optimal control of PDEs and ResNets.

Framework uses deep learning and statistical models to solve PDEs with discontinuous coefficients.

New method optimizes SDE models using continuous-time gradient descent.

We give new estimates for a critical elliptic system introduced by Rivière-Struwe in \cite{riviere_struwe} (see also the work of Rupflin \cite{rupflin} and Schikorra \cite{schikorra_frames}), which generalises PDE solved by harmonic (and almost harmonic) maps from a Euclidean ball $B_1 \In \R^n$ into Riemannian manifol…

Generative models improve inverse problems by providing tailored priors.

The paper develops a new approach to solve vector-valued PDEs on manifolds with minimal regularity.

We consider conformal immersions $f: T^2\rightarrow \mathbb{R}^3$ with the property that $H^2 f^*g_{\mathbb{R}^3}$ is a flat metric. These so called Dirac tori have the property that its Willmore energy is uniformly distributed over the surface and can be obtained using spin transformations of the plane by eigenvectors…

A new method infers parameters from PDEs using Gaussian processes.

In [29], Plebanski reformulated the anti-self-dual Einstein equations with non-zero scalar curvature as a first order PDE for a connection in an SO(3)-bundle over the four-manifold. The aim of this article is to place this differential equation in a new framework, in which it is both elliptic and a stationary point of …

Study anisotropic flows without global terms and solve dual Orlicz Christoffel-Minkowski problems.

Neural-PDE learns PDEs from data using LSTM, outperforming traditional methods.

Modeling wind dynamics in Saudi Arabia using deep learning and stochastic PDEs.

Given a Hermitian line bundle $L\to M$ over a closed, oriented Riemannian manifold $M$, we study the asymptotic behavior, as $ε\to 0$, of couples $(u_ε,\nabla_ε)$ critical for the rescalings \begin{align*} &E_ε(u,\nabla)=\int_M\Big(|\nabla u|^2+ε^2|F_\nabla|^2+\frac{1}{4ε^2}(1-|u|^2)^2\Big) \end{align*} of the self-dua…

We study the asymptotic behavior of solutions to the second boundary value problem for a parabolic PDE of Monge-Ampère type arising from optimal mass transport. Our main result is an exponential rate of convergence for solutions of this evolution equation to the stationary solution of the optimal transport problem. We …

Neural Q-learning tackles high-dimensional PDEs.

Partial differential equations (PDEs) are commonly derived based on empirical observations. However, recent advances of technology enable us to collect and store massive amount of data, which offers new opportunities for data-driven discovery of PDEs. In this paper, we propose a new deep neural network, called PDE-Net …

PDE-DKL combines NNs and GPs for high-dimensional PDE problems.

Solves second-order PDEs using quotients and differential invariants.

PRISMA uses PDE residuals for fast, robust, and accurate inference.

Meta-learning base distributions for efficient PDE solutions.

Kernel method learns PDEs from noisy data.

Solves a PDE for Landsberg surfaces using new Finsler surface insights.

Survey on conservation laws for geometric PDEs.

Using the theory of the symmetry group for PDEs [15, 17], we derive the symmetry group G associated to surfaces PDE. Several group invariant solutions of the surfaces PDE are given by solving a reduced system of partial differential equations.

Develops arithmetic PDE geometry concepts like curvature and cohomology.

Unified framework solves nonlinear PDEs and IPs using Gaussian processes.

Monotonicity formulae play a crucial role for many geometric PDEs, especially for their regularity theories. For minimal submanifolds in a Euclidean ball, the classical monotonicity formula implies that if such a submanifold passes through the centre of the ball, then its area is at least that of the equatorial disk. R…

The paper presents a PDE method for xVA incorporation in financial derivatives.

New PDEs of mixed type emerge in fluid mechanics and geometry.

We present a PDE-based framework that generalizes Group equivariant Convolutional Neural Networks (G-CNNs). In this framework, a network layer is seen as a set of PDE-solvers where geometrically meaningful PDE-coefficients become the layer's trainable weights. Formulating our PDEs on homogeneous spaces allows these net…

Automated PDE discovery from multiple noisy experiments.

In this paper we investigate the flow of surfaces by a class of symmetric functions of the principal curvatures with a mixed volume constraint. We consider compact surfaces without boundary that can be written as a graph over a sphere. The linearisation of the resulting fully nonlinear PDE is used to prove a short time…

Improved neural PDEs trained on augmented data enhance model accuracy and efficiency.

In recent years, data-driven methods have been developed to learn dynamical systems and partial differential equations (PDE). The goal of such work is discovering unknown physics and the corresponding equations. However, prior to achieving this goal, major challenges remain to be resolved, including learning PDE under …

Meta-learning neural networks to solve diverse PDEs efficiently.

It is shown that the characteristic vector field associated to a first order PDE has the same form of an infinitesimal generator of an odd-symplectic transformation with contact Hamiltonian the given PDE. It is considered under which condition such PDE has a characteristic vector field commuting with a generator of an …

In this paper, we present an initial attempt to learn evolution PDEs from data. Inspired by the latest development of neural network designs in deep learning, we propose a new feed-forward deep network, called PDE-Net, to fulfill two objectives at the same time: to accurately predict dynamics of complex systems and to …