← Back to all news
arXiv Machine Learning By Eunhan Ka, Ludovic Leclercq, Satish V. Ukkusuri

Observation-Aligned Two-Stage Domain Decomposition for Physics-Informed Traffic State Estimation with Sparse Fixed Sensors

Read the original on arXiv Machine Learning →

The paper introduces Observation‑Aligned Two‑Stage Domain Decomposition Physics‑Informed Neural Networks (TSDD‑PINN) for reconstructing traffic speed fields from sparse fixed sensors. It first trains a global PINN, then uses its residuals to partition the domain and warm‑start child networks, allowing spatial, temporal, or space‑time refinement. Experiments on the I‑24 MOTION dataset show that TSDD‑PINN achieves lower relative L2 error in most configurations and trains faster than the XPINN baseline, with performance depending on sensing density.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Machine Learning
Jun 5

On the training of physics-informed neural operators for solving parametric partial differential equations

arXiv:2606. 06164v1 Announce Type: new Abstract: Physics-informed neural operators (PINOs) aim to learn solution operators for partial differential equations by using the governing physics as supervision, rather than relying solely on paired input-output simulation data.

By Nanxi Chen, Chuanjie Cui, Airong Chen, Sifan Wang, Rujin Ma
llmscomputer-visionbenchmarks
More like this →
arXiv Machine Learning
Sep 15

Linearized PINN with pretrained nonlinear layers

The paper introduces a linearized Physics-Informed Neural Network (lPINN), a reduced‑order neural basis approach for solving forward and inverse differential equations. In an offline phase, lPINN learns continuous, differentiable neural basis functions from numerical solutions, which are then frozen for new problem instances; the online solution is obtained by minimizing the governing‑equation residual with additional constraints. Experiments on advection‑diffusion, Burgers', and nonlinear pendulum equations show that lPINN achieves lower solution and parameter errors than vanilla PINNs while reducing online inference times by up to three orders of magnitude, and its continuous representation generalizes to finer meshes without retraining.

By Wenhao Chen, Alexandre M. Tartakovsky
diffusion
More like this →