arXiv subjects

Weifan Liu

Publications and source records attributed to Weifan Liu.

5 recordsLinked to original sources

arXiv2026

Accelerating Microswimmer Simulations via a Heterogeneous Pipelined Parallel-in-Time Framework

Simulating large-scale microswimmer dynamics in viscous fluid poses significant challenges due to the coupled high spatial and temporal complexity. Conventional high-performance computing (HPC) methods often address these two dimensions in isolation, leaving a critical gap for synergistic acceleration. This paper introduces a heterogeneous CPU--GPU computing framework specifically optimized for the long-time simulation of filamentous microswimmers in viscous fluid. We propose a two-level parallelization strategy: (1) high-intensity GPU kernels to resolve the quadratic spatial interactions given by the Method of Regularized Stokeslets (MRS), and (2) a distributed MPI-GPU pipelined Parareal architecture to exploit temporal concurrency. By mapping the asynchronous pipeline onto multiple GPU devices, our framework effectively overlaps coarse and fine propagators, overcoming the serial bottlenecks of traditional Parareal method. Furthermore, we employ a GPU-optimized numerical routine for computing the matrix square root arising in the numerical scheme of the filamentous microswimmer simulations. Theoretical analysis of the efficiency improvement of the pipelined Parareal is presented. Numerical experiments demonstrate that the proposed framework achieves order-of-magnitude speedups over CPU-only methods, providing a scalable pathway for simulating complex emergent behaviors in large-scale biology and physics systems.

cs.DC↗

arXiv2025

FedGSCA: Medical Federated Learning with Global Sample Selector and Client Adaptive Adjuster under Label Noise

Federated Learning (FL) emerged as a solution for collaborative medical image classification while preserving data privacy. However, label noise, which arises from inter-institutional data variability, can cause training instability and degrade model performance. Existing FL methods struggle with noise heterogeneity and the imbalance in medical data. Motivated by these challenges, we propose FedGSCA, a novel framework for enhancing robustness in noisy medical FL. FedGSCA introduces a Global Sample Selector that aggregates noise knowledge from all clients, effectively addressing noise heterogeneity and improving global model stability. Furthermore, we develop a Client Adaptive Adjustment (CAA) mechanism that combines adaptive threshold pseudo-label generation and Robust Credal Labeling Loss. CAA dynamically adjusts to class distributions, ensuring the inclusion of minority samples and carefully managing noisy labels by considering multiple plausible labels. This dual approach mitigates the impact of noisy data and prevents overfitting during local training, which improves the generalizability of the model. We evaluate FedGSCA on one real-world colon slides dataset and two synthetic medical datasets under various noise conditions, including symmetric, asymmetric, extreme, and heterogeneous types. The results show that FedGSCA outperforms the state-of-the-art methods, excelling in extreme and heterogeneous noise scenarios. Moreover, FedGSCA demonstrates significant advantages in improving model stability and handling complex noise, making it well-suited for real-world medical federated learning scenarios.

cs.LG↗

arXiv2025

A parallel-in-time method based on the Parareal algorithm and High-Order Dynamic Mode Decomposition with applications to fluid simulations

The high cost of sequential time integration is one major constraint that limits the speedup of a time-parallel algorithm like the Parareal algorithm due to the difficulty of coarsening time steps in a stiff numerical problem. To address this challenge, we develop a parallel-in-time approach based on the Parareal algorithm, in which we construct a novel coarse solver using a data-driven method based on Dynamic Mode Decomposition in place of a classic time marching scheme. The proposed solver computes an approximation of the solution using two numerical schemes of different accuracies in parallel, and apply High-Order Dynamic Mode Decomposition (HODMD) to reduce the cost of sequential computations. Compared to the original Parareal algorithm, the proposed approach allows for the construction of low-cost coarse solvers for many complicated stiff problems. We demonstrate through several numerical examples in fluid dynamics that the proposed method can effectively reduce the serial computation cost and improve the parallel speedup of long-time simulations which are hard to accelerate using the original Parareal algorithm.

physics.comp-ph↗

arXiv2024

Mixing due to Solution-switch Limits the Performance of Electro-sorption for Desalination

Electro-sorption (ES) is a research frontier in electrochemical separation, with proven potential applications in desalination, wastewater treatment, and selective resource extraction. However, due to the limited adsorption capacity of film electrodes, ES requires short circuiting or circuit reversal, accompanied by solution-switch between the feed solution and receiving solution, to sustain desalination over many charging-discharge cycles. In the literature, solution-switch have been commonly ignored to simplify experimental procedures and its impacts on separation performance are thus not well understood. This study aims to provide a quantitative analysis of the impacts of mixing due to solution-switch on the performance of ES-based desalination. A numerical model of ES has been employed to evaluate the adverse effects of solution-switch on the desalination performance in three different operation modes. The analysis reveals that the impacts of mixing due to solution-switch are more severe with a larger concentration difference between the desalinated water and the brine and provides insights on the effectiveness of increasing electrode loading or specific capacity in mitigating the detrimental impacts of mixing. Even with state-of-the-art systems, producing freshwater from seawater or even brackish water with medium-to-high salinity is practically challenging due to the presence of solution-switch.

physics.chem-ph↗

arXiv2020

Steady-states of thin film droplets on chemically heterogeneous substrates

We study steady-state thin films on a chemically heterogeneous substrates of finite size, subject to no-flux boundary conditions. Based on the structure of the bifurcation diagram, we classify the one-dimensional steady-state solutions that exist on such substrates into six different branches and develop asymptotic estimates for the steady-states on each branch. We show using perturbation expansions, that leading order solutions provide good predictions of the steady-state thin films on stepwise-patterned substrates. The analysis in one dimension can be extended to axisymmetric solutions. We also examine the influence of the wettability contrast on linear stability and dynamics. Results are also applied to describe two-dimensional droplets on hydrophilic square patches and striped regions used in microfluidic applications.

physics.flu-dyn↗