Neural Relational Inference for Interacting Systems
This addresses the challenge of modeling interacting systems in physics and social dynamics, offering an unsupervised approach that is incremental by building on variational auto-encoders and graph neural networks.
The paper tackles the problem of inferring interactions in complex systems from observational data, introducing the neural relational inference (NRI) model, which accurately recovers ground-truth interactions in simulated physical systems and predicts dynamics in real-world data like motion capture.
Interacting systems are prevalent in nature, from dynamical systems in physics to complex societal dynamics. The interplay of components can give rise to complex behavior, which can often be explained using a simple model of the system's constituent parts. In this work, we introduce the neural relational inference (NRI) model: an unsupervised model that learns to infer interactions while simultaneously learning the dynamics purely from observational data. Our model takes the form of a variational auto-encoder, in which the latent code represents the underlying interaction graph and the reconstruction is based on graph neural networks. In experiments on simulated physical systems, we show that our NRI model can accurately recover ground-truth interactions in an unsupervised manner. We further demonstrate that we can find an interpretable structure and predict complex dynamics in real motion capture and sports tracking data.