Label Propagation Training Schemes for Physics-Informed Neural Networks and Gaussian Processes

arXiv:2404.05817v1 Announce Type: new
Abstract: This paper proposes a semi-supervised methodology for training physics-informed machine learning methods. This includes self-training of physics-informed neural networks and physics-informed Gaussian processes in isolation, and the integration of the two via co-training. We demonstrate via extensive numerical experiments how these methods can ameliorate the issue of propagating information forward in time, which is a common failure mode of physics-informed machine learning.

abstract arxiv cs.lg gaussian processes integration machine machine learning methodology networks neural networks numerical paper physics physics-informed processes propagation self-training semi-supervised training type via