Xinliang Liu

Yuelin Wang, Kai Yi, Xinliang Liu et al.

Neural message passing is a basic feature extraction unit for graph-structured data considering neighboring node features in network propagation from one layer to the next. We model such process by an interacting particle system with attractive and repulsive forces and the Allen-Cahn force arising in the modeling of phase transition. The dynamics of the system is a reaction-diffusion process which can separate particles without blowing up. This induces an Allen-Cahn message passing (ACMP) for graph neural networks where the numerical iteration for the particle system solution constitutes the message passing propagation. ACMP which has a simple implementation with a neural ODE solver can propel the network depth up to one hundred of layers with theoretically proven strictly positive lower bound of the Dirichlet energy. It thus provides a deep model of GNNs circumventing the common GNN problem of oversmoothing. GNNs with ACMP achieve state of the art performance for real-world node classification tasks on both homophilic and heterophilic datasets. Codes are available at https://github.com/ykiiiiii/ACMP.

Xinliang Liu, Bo Xu, Shuhao Cao et al.

Neural operators have emerged as a powerful tool for learning the mapping between infinite-dimensional parameter and solution spaces of partial differential equations (PDEs). In this work, we focus on multiscale PDEs that have important applications such as reservoir modeling and turbulence prediction. We demonstrate that for such PDEs, the spectral bias towards low-frequency components presents a significant challenge for existing neural operators. To address this challenge, we propose a hierarchical attention neural operator (HANO) inspired by the hierarchical matrix approach. HANO features a scale-adaptive interaction range and self-attentions over a hierarchy of levels, enabling nested feature computation with controllable linear cost and encoding/decoding of multiscale solution space. We also incorporate an empirical $H^1$ loss function to enhance the learning of high-frequency components. Our numerical experiments demonstrate that HANO outperforms state-of-the-art (SOTA) methods for representative multiscale problems.

Juncai He, Xinliang Liu, Jinchao Xu

In this work, we propose a concise neural operator architecture for operator learning. Drawing an analogy with a conventional fully connected neural network, we define the neural operator as follows: the output of the $i$-th neuron in a nonlinear operator layer is defined by $O_i(u) = σ\left( \sum_j W_{ij} u + B_{ij}\right)$. Here, $ W_{ij}$ denotes the bounded linear operator connecting $j$-th input neuron to $i$-th output neuron, and the bias $ B_{ij}$ takes the form of a function rather than a scalar. Given its new universal approximation property, the efficient parameterization of the bounded linear operators between two neurons (Banach spaces) plays a critical role. As a result, we introduce MgNO, utilizing multigrid structures to parameterize these linear operators between neurons. This approach offers both mathematical rigor and practical expressivity. Additionally, MgNO obviates the need for conventional lifting and projecting operators typically required in previous neural operators. Moreover, it seamlessly accommodates diverse boundary conditions. Our empirical observations reveal that MgNO exhibits superior ease of training compared to other CNN-based models, while also displaying a reduced susceptibility to overfitting when contrasted with spectral-type neural operators. We demonstrate the efficiency and accuracy of our method with consistently state-of-the-art performance on different types of partial differential equations (PDEs).

Xinliang Liu, Bingxin Zhou, Chutian Zhang et al.

Graph neural networks (GNNs) have achieved champion in wide applications. Neural message passing is a typical key module for feature propagation by aggregating neighboring features. In this work, we propose a new message passing based on multiscale framelet transforms, called Framelet Message Passing. Different from traditional spatial methods, it integrates framelet representation of neighbor nodes from multiple hops away in node message update. We also propose a continuous message passing using neural ODE solvers. It turns both discrete and continuous cases can provably achieve network stability and limit oversmoothing due to the multiscale property of framelets. Numerical experiments on real graph datasets show that the continuous version of the framelet message passing significantly outperforms existing methods when learning heterogeneous graphs and achieves state-of-the-art performance on classic node classification tasks with low computational costs.

Bo Xu, Xinliang Liu, Lei Zhang

This paper introduces a data-driven operator learning method for multiscale partial differential equations, with a particular emphasis on preserving high-frequency information. Drawing inspiration from the representation of multiscale parameterized solutions as a combination of low-rank global bases (such as low-frequency Fourier modes) and localized bases over coarse patches (analogous to dilated convolution), we propose the Dilated Convolutional Neural Operator (DCNO). The DCNO architecture effectively captures both high-frequency and low-frequency features while maintaining a low computational cost through a combination of convolution and Fourier layers. We conduct experiments to evaluate the performance of DCNO on various datasets, including the multiscale elliptic equation, its inverse problem, Navier-Stokes equation, and Helmholtz equation. We show that DCNO strikes an optimal balance between accuracy and computational cost and offers a promising solution for multiscale operator learning.

Juncai He, Xinliang Liu, Zitong Tian

This paper studies the numerical approximation of divergence-free vector fields by linearized shallow neural networks, also referred to as random feature models or finite neuron spaces. Combining the stable potential lifting for divergence-free fields with the scalar Sobolev integral representation theory via ReLU$^k$ networks, we derive a core integral representation of divergence-free Sobolev vector fields through antisymmetric potentials parameterized by linearized ReLU$^k$ neural networks. This representation, together with a quasi-uniform distribution argument for the inner parameters, yields optimal approximation rates for such linearized ReLU$^k$ neural networks under an exact divergence-free constraint. Numerical experiments in two and three spatial dimensions, including $L^2$ projection and steady Stokes problems, confirm the theoretical rates and illustrate the effectiveness of exactly divergence-free conditions in computation.

Xiang Cao, Qiaoqiao Ding, Xinliang Liu et al.

Ultrasound Computed Tomography (USCT) constitutes a nonlinear inverse problem with inherent ill-posedness that can benefit from regularization through diffusion generative priors. However, traditional approaches for solving Helmholtz equation-constrained USCT face three fundamental challenges when integrating these priors: PDE-constrained gradient computation, discretization-induced approximation errors, and computational imbalance between neural networks and numerical PDE solvers. In this work, we introduce \textbf{Diff-ANO} (\textbf{Diff}usion-based Models with \textbf{A}djoint \textbf{N}eural \textbf{O}perators), a novel framework that combines conditional consistency models with adjoint operator learning to address these limitations. Our two key innovations include: (1) a \textit{conditional consistency model} that enables measurement-conditional few-step sampling by directly learning a self-consistent mapping from diffusion trajectories, and (2) an \textit{adjoint operator learning} module that replaces traditional PDE solvers with neural operator surrogates for efficient adjoint-based gradient computation. To enable practical deployment, we introduce the batch-based Convergent Born Series (BCBS)--a memory-efficient strategy for online generation of neural operator training pairs. Comprehensive experiments demonstrate that Diff-ANO significantly improves both computational efficiency and reconstruction quality, especially under sparse-view and partial-view measurement scenarios.

Xinliang Liu, Tong Mao, Jinchao Xu

We present an estimation of the condition numbers of the \emph{mass} and \emph{stiffness} matrices arising from shallow ReLU$^k$ neural networks defined on the unit sphere~$\mathbb{S}^d$. In particular, when $\{θ_j^*\}_{j=1}^n \subset \mathbb{S}^d$ is \emph{antipodally quasi-uniform}, the condition number is sharp. Indeed, in this case, we obtain sharp asymptotic estimates for the full spectrum of eigenvalues and characterize the structure of the corresponding eigenspaces, showing that the smallest eigenvalues are associated with an eigenbasis of low-degree polynomials while the largest eigenvalues are linked to high-degree polynomials. This spectral analysis establishes a precise correspondence between the approximation power of the network and its numerical stability.

Juntao Wang, Xinliang Liu, Jiwei Jia

Loss functions for learning-based PDE preconditioners implicitly choose a \emph{metric} in which residuals are matched, yet most approaches still optimize an unpreconditioned Euclidean residual norm. For indefinite operators such as the high-frequency Helmholtz equation, this default metric can make both learning and iterative correction overly sensitive to near-resonant spectral components, while classical preconditioning succeeds precisely by reshaping the residual geometry. We show that the Born Series and shifted-Laplacian left preconditioning are linked by the identity $ I-G_ηV_η= G_ηA = L_η^{-1}A, $ which turns the reference Green operator $G_η$ into a natural Riesz-map residual metric $ R_η= G_η^\ast G_η$ and suggests measuring the physical residual via $ \|r\|_{R_η}=\|G_ηr\|_2. $ Building on this viewpoint, we propose a \emph{Neural Preconditioned Born Series} (NPBS) iteration that replaces the scalar CBS relaxation with a residual-driven neural operator, together with a metric-matched Born-series-inspired loss $\mathcal{L}_{\mathrm{bs}}^{R_η}$. The framework is architecture-agnostic and supports fast $\mathcal{O}(N\log N)$ evaluation via FFT/DST/DCT. Numerical experiments on heterogeneous Helmholtz problems demonstrate the effectiveness of our method, and its advantage becomes more pronounced as the systems grow more ill-conditioned; we then extend the framework to other PDE classes, including convection--diffusion--reaction equations and linearized Newton systems for nonlinear PDEs, where it also yields substantial iteration reductions.

Qinchen Song, Xinliang Liu, Lei Zhang

Conventional numerical solvers for the radiative transfer equation (RTE) exhibit severe sensitivity to medium parameters. To address this, we propose an operator learning framework that approximates the RTE solution map as a function of material properties. The core architecture, MgNet, preserves the solution operator framework established by recursive skeleton factorization (RSF) but substitutes its coefficient-specific sub-operators (e.g. smoother, prolongation operator and restriction operator) with learnable neural components. This design transcends the the fixed parametric structure of classical schemes, enabling data-driven sub-operator optimization and learning of their medium-parameter dependence. To mitigate spectral bias in operator learning, we introduce an adaptive angular compression technique within the loss function that dynamically suppresses high-frequency modes responsible for training instability. Comprehensive benchmarks demonstrate that, when deployed as a learned preconditioner, MgNet achieves at least 10 times acceleration over conventional preconditioners in the diffusive regime and maintains robust generalization to unseen parameter configurations. By unifying multilevel factorization structure with deep operator learning, this work establishes a physics-constrained operator-learning paradigm for radiative transport simulations.

Wenrui Hao, Xinliang Liu, Yahong Yang

Solving nonlinear partial differential equations (PDEs) with multiple solutions using neural networks has found widespread applications in various fields such as physics, biology, and engineering. However, classical neural network methods for solving nonlinear PDEs, such as Physics-Informed Neural Networks (PINN), Deep Ritz methods, and DeepONet, often encounter challenges when confronted with the presence of multiple solutions inherent in the nonlinear problem. These methods may encounter ill-posedness issues. In this paper, we propose a novel approach called the Newton Informed Neural Operator, which builds upon existing neural network techniques to tackle nonlinearities. Our method combines classical Newton methods, addressing well-posed problems, and efficiently learns multiple solutions in a single learning process while requiring fewer supervised data points compared to existing neural network methods.

Zhijun Zeng, Youjia Zheng, Hao Hu et al.

Accurate and efficient simulation of wave equations is crucial in computational wave imaging applications, such as ultrasound computed tomography (USCT), which reconstructs tissue material properties from observed scattered waves. Traditional numerical solvers for wave equations are computationally intensive and often unstable, limiting their practical applications for quasi-real-time image reconstruction. Neural operators offer an innovative approach by accelerating PDE solving using neural networks; however, their effectiveness in realistic imaging is limited because existing datasets oversimplify real-world complexity. In this paper, we present OpenBreastUS, a large-scale wave equation dataset designed to bridge the gap between theoretical equations and practical imaging applications. OpenBreastUS includes 8,000 anatomically realistic human breast phantoms and over 16 million frequency-domain wave simulations using real USCT configurations. It enables a comprehensive benchmarking of popular neural operators for both forward simulation and inverse imaging tasks, allowing analysis of their performance, scalability, and generalization capabilities. By offering a realistic and extensive dataset, OpenBreastUS not only serves as a platform for developing innovative neural PDE solvers but also facilitates their deployment in real-world medical imaging problems. For the first time, we demonstrate efficient in vivo imaging of the human breast using neural operator solvers.

Juncai He, Xinliang Liu, Jinchao Xu

In this work, we propose a novel framework to enhance the efficiency and accuracy of neural operators through self-composition, offering both theoretical guarantees and practical benefits. Inspired by iterative methods in solving numerical partial differential equations (PDEs), we design a specific neural operator by repeatedly applying a single neural operator block, we progressively deepen the model without explicitly adding new blocks, improving the model's capacity. To train these models efficiently, we introduce an adaptive train-and-unroll approach, where the depth of the neural operator is gradually increased during training. This approach reveals an accuracy scaling law with model depth and offers significant computational savings through our adaptive training strategy. Our architecture achieves state-of-the-art (SOTA) performance on standard benchmarks. We further demonstrate its efficacy on a challenging high-frequency ultrasound computed tomography (USCT) problem, where a multigrid-inspired backbone enables superior performance in resolving complex wave phenomena. The proposed framework provides a computationally tractable, accurate, and scalable solution for large-scale data-driven scientific machine learning applications.

Hao Wu, Yuan Gao, Ruijian Gou et al.

Reliable long-term forecasting of Earth system dynamics is fundamentally limited by instabilities in current artificial intelligence (AI) models during extended autoregressive simulations. These failures often originate from inherent spectral bias, leading to inadequate representation of critical high-frequency, small-scale processes and subsequent uncontrolled error amplification. Inspired by the nested grids in numerical models used to resolve small scales, we present TritonCast. At the core of its design is a dedicated latent dynamical core, which ensures the long-term stability of the macro-evolution at a coarse scale. An outer structure then fuses this stable trend with fine-grained local details. This design effectively mitigates the spectral bias caused by cross-scale interactions. In atmospheric science, it achieves state-of-the-art accuracy on the WeatherBench 2 benchmark while demonstrating exceptional long-term stability: executing year-long autoregressive global forecasts and completing multi-year climate simulations that span the entire available $2500$-day test period without drift. In oceanography, it extends skillful eddy forecast to $120$ days and exhibits unprecedented zero-shot cross-resolution generalization. Ablation studies reveal that this performance stems from the synergistic interplay of the architecture's core components. TritonCast thus offers a promising pathway towards a new generation of trustworthy, AI-driven simulations. This significant advance has the potential to accelerate discovery in climate and Earth system science, enabling more reliable long-term forecasting and deeper insights into complex geophysical dynamics.

Xinliang Liu, Xia Yang, Chen-Song Zhang et al.

This research investigates the application of Multigrid Neural Operator (MgNO), a neural operator architecture inspired by multigrid methods, in the simulation for multiphase flow within porous media. The architecture is adjusted to manage a variety of crucial factors, such as permeability and porosity heterogeneity. The study extendes MgNO to time-dependent porous media flow problems and validate its accuracy in predicting essential aspects of multiphase flows. Furthermore, the research provides a detailed comparison between MgNO and Fourier Neural Opeartor (FNO), which is one of the most popular neural operator methods, on their performance regarding prediction error accumulation over time. This aspect provides valuable insights into the models' long-term predictive stability and reliability. The study demonstrates MgNO's capability to effectively simulate multiphase flow problems, offering considerable time savings compared to traditional simulation methods, marking an advancement in integrating data-driven methodologies in geoscience applications.

Bingxin Zhou, Xinliang Liu, Yuehua Liu et al.

Many real-world relational systems, such as social networks and biological systems, contain dynamic interactions. When learning dynamic graph representation, it is essential to employ sequential temporal information and geometric structure. Mainstream work achieves topological embedding via message passing networks (e.g., GCN, GAT). The temporal evolution, on the other hand, is conventionally expressed via memory units (e.g., LSTM or GRU) that possess convenient information filtration in a gate mechanism. Though, such a design prevents large-scale input sequence due to the over-complicated encoding. This work learns from the philosophy of self-attention and proposes an efficient spectral-based neural unit that employs informative long-range temporal interaction. The developed spectral window unit (SWINIT) model predicts scalable dynamic graphs with assured efficiency. The architecture is assembled with a few simple effective computational blocks that constitute randomized SVD, MLP, and graph Framelet convolution. The SVD plus MLP module encodes the long-short-term feature evolution of the dynamic graph events. A fast framelet graph transform in the framelet convolution embeds the structural dynamics. Both strategies enhance the model's ability on scalable analysis. In particular, the iterative SVD approximation shrinks the computational complexity of attention to O(Nd\log(d)) for the dynamic graph with N edges and d edge features, and the multiscale transform of framelet convolution allows sufficient scalability in the network training. Our SWINIT achieves state-of-the-art performance on a variety of online continuous-time dynamic graph learning tasks, while compared to baseline methods, the number of its learnable parameters reduces by up to seven times.