Unsupervised learning of latent physical properties using perception-prediction networks
Author(s)Zheng, David Y.; Wu, Jiajun; Tenenbaum, Joshua B
MetadataShow full item record
We propose a framework for the completely unsupervised learning of latent object properties from their interactions: the perception-prediction network (PPN). Consisting of a perception module that extracts representations of latent object properties and a prediction module that uses those extracted properties to simulate system dynamics, the PPN can be trained in an end-to-end fashion purely from samples of object dynamics. The representations of latent object properties learned by PPNs not only are sufficient to accurately simulate the dynamics of systems comprised of previously unseen objects, but also can be translated directly into human-interpretable properties (e.g. mass, coefficient of restitution) in an entirely unsuper-vised manner. Crucially, PPNs also generalize to novel scenarios: their gradient-based training can be applied to many dynamical systems and their graph-based structure functions over systems comprised of different numbers of objects. Our results demonstrate the efficacy of graph-based neural architectures in object-centric inference and prediction tasks, and our model has the potential to discover relevant object properties in systems that are not yet well understood.
DepartmentMassachusetts Institute of Technology. Department of Electrical Engineering and Computer Science; Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences
Association For Uncertainty in Artificial Intelligence (AUAI)
Zheng, David et al. “Unsupervised learning of latent physical properties using perception-prediction networks.” David Zheng's M. Eng. degree thesis, Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, June 2018, © 2018 The Author(s)
Author's final manuscript