arXiv subjects

Tangjun Wang

Publications and source records attributed to Tangjun Wang.

8 recordsLinked to original sources

arXiv2026

Weighted Laplacian Flow: A Deterministic Particle Flow with Provable Convergence

Sampling from a target probability density is a fundamental task in statistics, machine learning, and scientific computing. We introduce weighted Laplacian flow, a deterministic particle-flow method that transports samples from a tractable initial density to a target density known up to normalization. The method evolves the logarithmic density ratio between the target and the transported distribution and constructs the particle velocity by solving a weighted Poisson equation associated with the target density. This design avoids the need to choose a kernel and enables direct control of particle weights along the flow. We establish the global well-posedness of the proposed PDE system and prove that the transported density converges to the target density in both $L^\infty$ distance and Kullback-Leibler divergence. Under a sublinear forcing condition, the method achieves exact convergence in finite time. Numerical experiments on multimodal, heavy-tailed, and ten-dimensional targets demonstrate that weighted Laplacian flow can perform long-range mass transport, overcome energy barriers.

math.NA↗

arXiv2026

A Nonlocal $p$-Laplacian Interface Model with Sharp Interface

We propose an energy-based nonlocal $p$-Laplacian interface problem. Neumann interface conditions are naturally formulated via the energy, while Dirichlet conditions are enforced through a penalty term. A key feature is that the model retains a sharp interface, which facilitates extension to other interface problems; we illustrate this by developing a nonlocal approximation for the $p$-Laplacian interface problem with membrane conditions. By establishing $\Gamma$-convergence and compactness, we prove that as the nonlocal horizon vanishes, minimizers of the nonlocal functionals converge to those of the local counterparts. Numerical experiments using an efficient finite element method confirm the convergence.

math.AP↗

arXiv2025

A Nonlocal Biharmonic Model with $\Gamma$-Convergence to Local Model and an Efficient Numerical Method

Nonlocal models and their associated theories have been extensively investigated in recent years. Among these, nonlocal versions of the classical Laplace operator have attracted the most attention, while higher-order nonlocal operators have been studied far less. In this work, we focus on the nonlocal counterpart of the classical biharmonic operator together with the clamped boundary condition ($u$ and $\frac{\partial u}{\partial n}$ are given on the boundary). We develop the variational formulation of a nonlocal biharmonic model, establish the existence and uniqueness of its solution, and analyze its convergence as the nonlocal horizon approaches zero. In addition, we propose an efficient finite element method to solve the nonlocal model and the numerical results verify the analytical properties of the nonlocal model and its solution.

math.AP↗

arXiv2024

Interface Laplace Learning: Learnable Interface Term Helps Semi-Supervised Learning

We introduce a novel framework, called Interface Laplace learning, for graph-based semi-supervised learning. Motivated by the observation that an interface should exist between different classes where the function value is non-smooth, we introduce a Laplace learning model that incorporates an interface term. This model challenges the long-standing assumption that functions are smooth at all unlabeled points. In the proposed approach, we add an interface term to the Laplace learning model at the interface positions. We provide a practical algorithm to approximate the interface positions using k-hop neighborhood indices, and to learn the interface term from labeled data without artificial design. Our method is efficient and effective, and we present extensive experiments demonstrating that Interface Laplace learning achieves better performance than other recent semi-supervised learning approaches at extremely low label rates on the MNIST, FashionMNIST, and CIFAR-10 datasets.

cs.LG↗

arXiv2024

Convection-Diffusion Equation: A Theoretically Certified Framework for Neural Networks

In this paper, we study the partial differential equation models of neural networks. Neural network can be viewed as a map from a simple base model to a complicate function. Based on solid analysis, we show that this map can be formulated by a convection-diffusion equation. This theoretically certified framework gives mathematical foundation and more understanding of neural networks. Moreover, based on the convection-diffusion equation model, we design a novel network structure, which incorporates diffusion mechanism into network architecture. Extensive experiments on both benchmark datasets and real-world applications validate the performance of the proposed model.

cs.LG↗

arXiv2023

An axiomatized PDE model of deep neural networks

Inspired by the relation between deep neural network (DNN) and partial differential equations (PDEs), we study the general form of the PDE models of deep neural networks. To achieve this goal, we formulate DNN as an evolution operator from a simple base model. Based on several reasonable assumptions, we prove that the evolution operator is actually determined by convection-diffusion equation. This convection-diffusion equation model gives mathematical explanation for several effective networks. Moreover, we show that the convection-diffusion model improves the robustness and reduces the Rademacher complexity. Based on the convection-diffusion equation, we design a new training method for ResNets. Experiments validate the performance of the proposed method.

cs.LG↗

arXiv2023

A Nonlocal diffusion model with $H^1$ convergence for Dirichlet Boundary

In this paper, we present a nonlocal model for Poisson equation and corresponding eigenproblem with Dirichlet boundary condition. In the direct derivation of the nonlocal model, normal derivative is required which is not known for Dirichlet boundary. To overcome this difficulty, we treat the normal derivative as an auxiliary variable and derive corresponding nonlocal approximation of the boundary condition. For this specifically designed nonlocal mode, we can prove its well-posedness and convergence to the counterpart continuous model. The nonlocal model is carefully designed such that coercivity and symmetry are preserved. Based on these good properties, we can prove the nonlocal model converges with first order rate in $H^1$ norm. Our model can be naturally extended to Poisson problems with Robin boundary and corresponding eigenvalue problem.

math.AP↗

arXiv2021

Diffusion Mechanism in Residual Neural Network: Theory and Applications

Diffusion, a fundamental internal mechanism emerging in many physical processes, describes the interaction among different objects. In many learning tasks with limited training samples, the diffusion connects the labeled and unlabeled data points and is a critical component for achieving high classification accuracy. Many existing deep learning approaches directly impose the fusion loss when training neural networks. In this work, inspired by the convection-diffusion ordinary differential equations (ODEs), we propose a novel diffusion residual network (Diff-ResNet), internally introduces diffusion into the architectures of neural networks. Under the structured data assumption, it is proved that the proposed diffusion block can increase the distance-diameter ratio that improves the separability of inter-class points and reduces the distance among local intra-class points. Moreover, this property can be easily adopted by the residual networks for constructing the separable hyperplanes. Extensive experiments of synthetic binary classification, semi-supervised graph node classification and few-shot image classification in various datasets validate the effectiveness of the proposed method.

cs.LG↗