Differential equation and probability inspired graph neural networks for latent variable learning
This work addresses latent variable learning for machine learning applications, but it appears incremental as it combines existing tools like variational inference and differential equations with graph neural networks.
The paper tackles the problem of learning latent variables from observations by proposing graph neural networks inspired by probabilistic theory and differential equations, using variational inference and differential equations to solve subspace learning problems, but no concrete results or numbers are provided.
Probabilistic theory and differential equation are powerful tools for the interpretability and guidance of the design of machine learning models, especially for illuminating the mathematical motivation of learning latent variable from observation. Subspace learning maps high-dimensional features on low-dimensional subspace to capture efficient representation. Graphs are widely applied for modeling latent variable learning problems, and graph neural networks implement deep learning architectures on graphs. Inspired by probabilistic theory and differential equations, this paper conducts notes and proposals about graph neural networks to solve subspace learning problems by variational inference and differential equation.