Bond Graphs for multi-physics informed Neural Networks for multi-variate time series
This work addresses the problem of modeling complex multi-physical systems for researchers and practitioners in hybrid AI, though it appears incremental as it builds on existing physical-informed learning and graph neural network techniques.
The paper tackles the challenge of applying physical-informed machine learning to complex multi-physical and multi-domain phenomena, which existing methods are not adapted for, by proposing a Neural Bond graph Encoder (NBgE) that leverages Bond Graphs and Message Passing Graph Neural Networks to produce multi-physics-informed representations; experiments on a Direct Current Motor and the Respiratory System demonstrate its effectiveness on multivariate time-series forecasting.
In the trend of hybrid Artificial Intelligence techniques, Physical-Informed Machine Learning has seen a growing interest. It operates mainly by imposing data, learning, or architecture bias with simulation data, Partial Differential Equations, or equivariance and invariance properties. While it has shown great success on tasks involving one physical domain, such as fluid dynamics, existing methods are not adapted to tasks with complex multi-physical and multi-domain phenomena. In addition, it is mainly formulated as an end-to-end learning scheme. To address these challenges, we propose to leverage Bond Graphs, a multi-physics modeling approach, together with Message Passing Graph Neural Networks. We propose a Neural Bond graph Encoder (NBgE) producing multi-physics-informed representations that can be fed into any task-specific model. It provides a unified way to integrate both data and architecture biases in deep learning. Our experiments on two challenging multi-domain physical systems - a Direct Current Motor and the Respiratory System - demonstrate the effectiveness of our approach on a multivariate time-series forecasting task.