Tag
一篇发表于 Journal of the Brazilian Society of Mechanical Sciences and Engineering 的综述性论文，基于 COBEM 2025 圆桌讨论，系统探讨了 AI 在机械工程中的基础、机遇、挑战与责任，强调 AI 应与物理推理和工程严谨性结合而非取代工程判断。
提出SNTO框架，将状态、伴随和设计变量用神经网络联合表示并优化，结合多GPU硬件感知训练，实现可扩展的应力约束拓扑优化，可处理高达八百万配置点的三维大规模问题。
提出一种结构感知的核函数设计方法，用于高斯过程分布式端口哈密顿系统（GP-dPHS），通过张量积协方差解耦状态与空间依赖关系并融入空间异质性先验知识，实验表明在力场与轨迹重建上显著提升精度。
Anima Anandkumar contributed to the UN Secretary-General's Scientific Advisory Board brief on AI and climate modeling, highlighting progress from FourCastNet and the challenges of physics-informed AI climate prediction, data gaps, and equitable access.
Introduces Newmark-β-DGN, a graph neural network framework that infers internal mechanical responses from coarse-step motion observations, enabling long-horizon prediction and analysis of forces without direct supervision.
This paper studies scaling laws for physics-aware surrogate models in AC optimal power flow, demonstrating that physics-informed training objectives improve constraint satisfaction at scale with different scaling rates than standard loss functions.
This paper introduces Response Renormalization, a backward-pass framework to improve training stability in Deep Equilibrium Models by addressing near-singular Jacobian issues, demonstrated across various multiphysics applications.
This paper studies when conservation laws can be certified in learned latent world models, proposing bounded horizons that guarantee how long rollouts stay on physical invariant level sets using measurable model defects.
This paper proposes PE-MHL, a Physics-Encoded Modular Hybrid Layer framework that incrementally refines a physics-based model with data-driven sub-models, providing theoretical convergence guarantees and outperforming monolithic networks on control benchmarks.
Researchers propose a Physics-Informed Machine Learning (PIML) framework that integrates hydrological constraints into an LSTM loss function to improve short-term flood forecasting, particularly in data-scarce regimes. A 'Trend Alignment' constraint enforcing consistency between precipitation and discharge trends improves Nash-Sutcliffe Efficiency and eliminates unphysical predictions during extreme events.
This paper presents a novel deep learning approach to predict inertial lift forces in microfluidic devices without explicit geometric parameters, enabling better generalization to unseen channel cross-sections compared to previous models.
This paper proposes an active learning framework to couple high-fidelity Modelica simulations with simpler surrogate models (SINDyC, FNN, GRU) for creating efficient digital twins of thermal energy distribution systems. The approach significantly reduces the number of simulation trajectories needed while maintaining predictive accuracy and enabling uncertainty quantification.