Corpus ID: 211133037

Learning Group Structure and Disentangled Representations of Dynamical Environments

@article{Quessard2020LearningGS,
  title={Learning Group Structure and Disentangled Representations of Dynamical Environments},
  author={Robin Quessard and Thomas D. Barrett and W. Clements},
  journal={ArXiv},
  year={2020},
  volume={abs/2002.06991}
}
Discovering the underlying structure of a dynamical environment involves learning representations that are interpretable and disentangled, which is a challenging task. In physics, interpretable representations of our universe and its underlying dynamics are formulated in terms of representations of groups of symmetry transformations. We propose a physics-inspired method, built upon the theory of group representation, that learns a representation of an environment structured around the… Expand

Figures from this paper

Addressing the Topological Defects of Disentanglement via Distributed Operators
A Metric for Linear Symmetry-Based Disentanglement
GEM: Group Enhanced Model for Learning Dynamical Control Systems
Disentangling by Subspace Diffusion
Linear Disentangled Representations and Unsupervised Action Estimation

References

SHOWING 1-10 OF 43 REFERENCES
Symmetry-Based Disentangled Representation Learning requires Interaction with Environments
Towards a Definition of Disentangled Representations
A Disentangled Recognition and Nonlinear Dynamics Model for Unsupervised Learning
Understanding disentangling in β-VAE
Disentangled State Space Representations
Understanding disentangling in $\beta$-VAE
...
1
2
3
4
5
...