Date/Time : 11 Sep 2024 16:30
Venue : S17 #06-11 (SR6)
Speaker : Yang Liu, National University of Singapore

Can we build a single large model for a wide range of PDE-related scientific learning tasks? Can this model generalize to new PDEs without any fine-tuning? In-context operator learning and the corresponding model In-Context Operator Networks (ICON) represent an initial exploration of these questions. We present a detailed methodology for solving PDE problems with ICON, and show how a single ICON model can make forward and reverse predictions for different equations with different strides, provided with appropriately designed data prompts. We show positive evidence for the questions above, through a study on 1D scalar nonlinear conservation laws, a family of PDEs with temporal evolution. In particular, we show that an ICON model trained on conservation laws with cubic flux functions can generalize well to some other flux functions of more general forms, without fine-tuning. We will also discuss different theoretical interpretations of in-context learning.

Add to calendar :

Date/Time : 04 Sep 2024 16:30
Venue : S17 #06-11 (SR6)
Speaker : Lin Bo, National University of Singapore

This talk focuses on the development of time-dependent regularized 13-moment equations for general elastic collision models within the linear regime. The detailed derivation demonstrates that these proposed equations exhibit super-Burnett order for small Knudsen numbers and feature a symmetric structure. A novel modification to stable boundary conditions is introduced to remove boundary layers. The effectiveness of the model is validated through numerical simulations of channel flows.

Add to calendar :

Date/Time : 28 Aug 2024 16:30
Venue : S17 #04-05 (SR2)
Speaker : Zhang Xiaolong, Nanyang Technological University & Hunan Normal University

The Logarithmic Schrödinger Equation (LogSE) exhibits rich dynamics and possesses unique properties compared to the standard Schrödinger equation with cubic nonlinearity. However, the logarithmic nonlinear term $f(u)=uln(∣u∣^2)$ presents significant challenges for both numerical solutions and error analysis, primarily due to low regularity and singularity. In this talk, we will characterize this low regularity within suitable fractional Sobolev spaces and derive the first set of error estimates for the time-splitting Fourier spectral methods with initial data of fractional-order regularity. These results are supported by recent regularity estimates for the LogSE from the PDE community and ample numerical evidence.(Joint work with Prof. Li-Lian Wang, NTU)

Add to calendar :

Date/Time : 08 May 2024 16:30
Venue : S17 #04-05 (SR2)
Speaker : Junfu Yao, Johns Hopkins University

Minimal surfaces are known as critical points to the area functional. In this talk, I’ll focus on the question of finding minimal surfaces with prescribed behavior at the asymptotic infinity. This is called the asymptotic Plateau problem. I’ll briefly review previous results and present a min-max theory for finding solutions to this problem. As a corollary of the mountain-pass theorem, I’ll also discuss a topological uniqueness result.

Add to calendar :

Date/Time : 20 Mar 2024 16:30
Venue : S17 #05-11 (SR5)
Speaker : Jin Zeyu, Peking University

A promising approach to investigating high-dimensional problems is to identify their intrinsically low-dimensional features, which can be achieved through recently developed techniques for effective low-dimensional representation of functions such as machine learning. Based on available finite-dimensional approximate solution manifolds, we propose a novel model reduction framework for kinetic equations. The method employs projections onto tangent bundles of approximate manifolds, naturally resulting in first-order hyperbolic systems. Under certain conditions on the approximate manifolds, the reduced models preserve several crucial properties, including hyperbolicity, conservation laws, entropy dissipation, finite propagation speed, and linear stability. For the first time, we rigorously discuss the relation between the H-theorem of kinetic equations and the linear stability conditions of reduced systems, determining the choice of Riemannian metrics involved in the model reduction. The framework is widely applicable for the model reduction of many models in kinetic theory. Joint work with Ruo Li.

Add to calendar :

Date/Time : 13 Mar 2024 16:30
Venue : S17 #05-11 (SR5)
Speaker : Lin Bo, National University of Singapore

The committor function is a central object for quantifying the transitions between metastable states of dynamical systems. Recently, a number of computational methods based on deep neural networks have been developed for computing the high-dimensional committor function. The success of the methods relies on sampling adequate data for the transition, which still is a challenging task for complex systems at low temperatures. This talk will present a deep learning method with two novel adaptive sampling schemes. In the two schemes, the data are generated actively from a modified potential where the bias potential is constructed from the learned committor function. We theoretically show that the data in the two schemes are uniformly distributed along the transition tube. The efficiency of the method is illustrated on high-dimensional systems including the alanine dipeptide and a dimer system immersed in the water box.

Add to calendar :

Date/Time : 06 Mar 2024 16:30
Venue : S17 #05-11 (SR5)
Speaker : Li Yifei, National University of Singapore

Designing a numerical scheme that can preserve the geometric structure for anisotropic surface diffusion with an arbitrary anisotropic surface energy is a long-standing problem. In this talk, we propose and analyze a structure-preserving parametric finite element methods (SP-PFEM) for the evolution of a closed curve in 2D, which preserve two geometric structures – area conservation and energy dissipation – at the full-discretized level, The SP-PFEM innovates with a novel surface energy matrix and the Cahn-Hoffman ξ-vector, leading to a new geometric identity for dealing with the weighted mean curvature. This new identity allows our SP-PFEM to be easily extended to various geometric flows with anisotropic effects. Extensive numerical results demonstrate its efficiency, stability, and success in other geometric flows.

Add to calendar :

Date/Time : 28 Feb 2024 16:30
Venue : S17 #05-12 (SR4)
Speaker : Guo Ruchi, The Chinese University of Hong Kong

In physics and mathematics, a large class of PDE systems can be formulated as energy minimization problems subject to specific constraints. Lagrange multipliers are widely used for solving such problems, which however leads to min-max optimization problems known as saddle point systems. The development of fast solvers for saddle point systems, especially those nonlinear ones, presents considerable challenges in the sense that

(i) one has to consider the preconditioning in the primal and dual directions collectively;

(ii) the preconditioners have to evolve in iterations to accommodate the nonlinearity;

(iii) all the subsystems and subproblems cannot be solved exactly due to the high computational cost involved.

In this work, we develop an efficient transformed primal-dual (TPD) algorithm, and this is done by modifying the original gradient flow from the perspective of matrix factorization. We prove the global optimal linear convergence in terms of the condition numbers measured by variable metrics. To demonstrate the effectiveness of our approach, we apply the algorithm to a Darcy-Forchheimer model and a nonlinear ferromagnetic model, showing it is much more efficient than some traditional methods.

Add to calendar :

Date/Time : 14 Feb 2024 16:30
Venue : S17 #04-06 (SR1)
Speaker : Liu Xin, Chinese Academy of Sciences

We propose constraint dissolving approaches for optimization problems over a class of Riemannian manifolds. In these proposed approaches, solving a Riemannian optimization problem is transferred into the unconstrained minimization of a constraint dissolving function named CDF. Different from existing exact penalty functions, the exact gradient and Hessian of CDF are easy to compute. We study the theoretical properties of CDF and prove that the original problem and CDF have the same first-order and second-order stationary points, local minimizers, and Łojasiewicz exponents in a neighborhood of the feasible region. Remarkably, the convergence properties of our proposed constraint dissolving approaches can be directly inherited from the existing rich results in unconstrained optimization. Therefore, the proposed constraint dissolving approaches build up short cuts from unconstrained optimization to Riemannian optimization. Several illustrative examples further demonstrate the potential of the proposed approaches.

Add to calendar :