Zhiyuan Ren

Dr. Zhiyuan Ren

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
21 July 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ZRZhiyuan Ren
Zhiyuan Ren
SZShijie Zhou
Shijie Zhou
DLDong Liu
Dong Liu
QLQihe Liu
Qihe Liu
Physics-informed neural networks (PINNs) have emerged as a transformative methodology integrating deep learning with scientific computing. This review establishes a three-dimensional analytical framework to systematically decode PINNs’ development through methodological innovation, theoretical breakthroughs, and cross-disciplinary convergence. The contributions include threefold: First, identifying the co-evolutionary path of algorithmic architectures from adaptive optimization (neural tangent kernel-guided weighting achieving 230% convergence acceleration in Navier-Stokes solutions) to hybrid numerical-deep learning integration (5× speedup via domain decomposition) and second, constructing bidirectional theory-application mappings where convergence analysis (operator approximation theory) and generalization guarantees (Bayesian-physical hybrid frameworks) directly inform engineering implementations, as validated by 72% cost reduction compared to FEM in high-dimensional spaces (p<0.01,n=15 benchmarks). Third, pioneering cross-domain knowledge transfer through application-specific architectures: TFE-PINN for turbulent flows (5.12±0.87% error in NASA hypersonic tests), ReconPINN for medical imaging (▵SSIM=+0.18±0.04 on multi-institutional MRI), and SeisPINN for seismic systems (0.52±0.18 km localization accuracy). We further present a technological roadmap highlighting three critical directions for PINN 2.0: neuro-symbolic, federated physics learning, and quantum-accelerated optimization. This work provides methodological guidelines and theoretical foundations for next-generation scientific machine learning systems.

14 December 2025 in Journal of Artificial Intelligence Research
Verified
Journal Article
Closed access
PDF file
ZRZhiyuan Ren
Zhiyuan Ren
SZShijie Zhou
Shijie Zhou
DLDong Liu
Dong Liu
QLQihe Liu
Qihe Liu
Background: Integral equations play a crucial role in modeling complex systems across various scientific disciplines. However, traditional numerical methods and existing physics-informed neural networks (PINNs) face substantial challenges, including the curse of dimensionality, uncontrolled error propagation, and limited generalization capabilities. Objectives: This paper aims to overcome these limitations by developing a robust and scalable solver for high-dimensional and nonlinear integral equations. The primary goal is to achieve higher accuracy and efficiency compared to traditional methods and existing deep learning approaches. Methods: We present the enhanced physics-informed neural network (EPINN), a novel framework that incorporates three key innovations: 1) a variable-order operator decomposition theory that transforms integral equations into well-posed differential systems, thereby mitigating error accumulation, 2) a differentiable primal function projection layer that ensures physical consistency within the Sobolev spaces, and 3) a boundary-aware multi-objective training paradigm that improves generalization. Results: Experimental validation across five benchmark cases spanning two to four dimensions, including linear/nonlinear Volterra/Fredholm and hybrid Volterra-Fredholm integral equations, demonstrates the superior performance of EPINN. Compared with traditional methods, EPINN reduces relative errors by 1 to 2 orders of magnitude, while achieving over 92% accuracy with limited training data. When compared with existing deep learning solvers, EPINN provides significant improvements in computational efficiency (with a speedup factor of 3 to 6 times) and accuracy (error reduction of 23% to 85%). Conclusions: These advancements establish EPINN as a robust and scalable solver for high-dimensional and nonlinear integral equations, with wide-ranging applications in computational physics and engineering. The success of EPINN suggests that integrating physical principles with neural networks can lead to substantial improvements in solving complex mathematical problems.

1 January 2025 in AIMS Mathematics
Verified
Journal Article
Open access
PDF file
ZRZhiyuan Ren
Zhiyuan Ren
DLDong Liu
Dong Liu
ZLZhen Liao
Zhen Liao
SZShijie Zhou
Shijie Zhou
QLQihe Liu
Qihe Liu
Solving integral equations via deep learning encounters significant computational bottlenecks when order-reduction techniques transform problems into strongly coupled differential systems requiring multi-network collaborative training. While achieving high accuracy, existing distributed training paradigms exhibit fundamental limitations. Data parallelism suffers from prohibitive gradient synchronization overhead in multi-network coupling scenarios, while pipeline parallelism struggles with bubble inefficiencies in the shallow architectures typical of scientific computing. To overcome these challenges, we proposed the computationally efficient parallel training framework (CEPTF), which introduces three key innovations. A unified computational efficiency metric balancing acceleration gains with resource costs, mathematical-aware dynamic partitioning that adapts to equation structure and hardware constraints, and hybrid parallelism integrating optimized communication topologies with constraint-preserving synchronization. Comprehensive validation across linear/nonlinear Fredholm/Volterra equations demonstrates that CEPTF achieves 3.18$ \times $ to 6.32$ \times $ acceleration (318.6%–632.9% speedup ratio) while maintaining solution accuracy of $ 10^{-7} $ to $ 10^{-9} $ magnitude, outperforming established parallel paradigms by 1.8$ \times $ to 3.2$ \times $ in speedup ratios and 42%–67% in computational efficiency. The framework's adaptability to heterogeneous computing environments and robust performance under challenging conditions, including singular kernels and irregular domains, establishes a new paradigm for scalable scientific machine learning.

15 October 2025 in Engineering Applications of Artificial Intelligence
Verified
Journal Article
Closed access
ZRZhiyuan Ren
Zhiyuan Ren
SZShijie Zhou
Shijie Zhou
DLDong Liu
Dong Liu
QLQihe Liu
Qihe Liu

1 January 2026 in AIMS Mathematics
Verified
Journal Article
Open access
PDF file
ZRZhiyuan Ren
Zhiyuan Ren
RYRuilong Yu
Ruilong Yu
YZYi Zeng
Yi Zeng
SZShijie Zhou
Shijie Zhou
QLQihe Liu
Qihe Liu
Deep learning has become a key approach for solving integral equations, a central theme in numerical computation. However, existing methods often struggle with precision and efficiency. This study presents the parallel variable-order iteration algorithm (P-VOI), inspired by the variable-order method in the R-AIM algorithm, specifically designed to resolve continuous integral equations using deep learning. The P-VOI algorithm transforms these equations into differential forms by applying the Newton-Leibniz formula. It employs parallel iterative processes, allowing neural networks to approximate the unknown and primitive functions simultaneously. This integration of parallel iteration significantly enhances computational efficiency compared to R-AIM. Additionally, the P-VOI algorithm uses neural networks as trial functions during training, addressing challenges that arise when integral kernels have zeros. This method reduces systematic errors from finite-sum representations, improving solution accuracy. Empirical evaluations demonstrate that the P-VOI algorithm achieves superior convergence accuracy (mean squared error in the range of $ 10^{-7} $ to $ 10^{-9} $) and efficiency, with speedups of 2.9× to 5.2× over R-AIM. Compared with non-variable-order methods, this corresponds to an improvement of approximately three orders of magnitude in accuracy based on MSE; relative to the direct variable-order baseline R-AIM, the accuracy remains of the same order of magnitude with slightly higher precision.

9 August 2026 in Axioms
Verified
Journal Article
Open access
PDF file
ZRZhiyuan Ren
Zhiyuan Ren
RYRuilong Yu
Ruilong Yu
YZYi Zeng
Yi Zeng
SZShijie Zhou
Shijie Zhou
This review surveys recent deep learning approaches for solving integral equations, categorizing them into three methodological families: physics-informed embedding, spectral/topological acceleration, and hybrid symbolic–numeric frameworks. The main findings are threefold. First, these methods achieve promising empirical accuracy in oscillatory, high-dimensional, and singular-kernel settings, yet their theoretical foundations remain largely incomplete. Second, from an axiomatic perspective, most approaches lack rigorous guarantees of convergence, stability, and spectral consistency; we formulate five testable propositions that a complete theory should satisfy. Third, we identify five specific unresolved theoretical questions and outline a focused research agenda toward a mathematically rigorous theory of neural operator approximation for integral equations. The novelty of this review lies in its dual computational–axiomatic evaluation and its provision of a structured, problem-oriented framework for future investigations.