q-VAE for Disentangled Representation Learning and Latent Dynamical Systems

@article{Kobayashi2020qVAEFD,
  title={q-VAE for Disentangled Representation Learning and Latent Dynamical Systems},
  author={Taisuke Kobayashi},
  journal={IEEE Robotics Autom. Lett.},
  year={2020},
  volume={5},
  pages={5669-5676}
}
  • Taisuke Kobayashi
  • Published 2020
  • Computer Science, Mathematics
  • IEEE Robotics Autom. Lett.
  • A variational autoencoder (VAE) derived from Tsallis statistics called q-VAE is proposed. In the proposed method, a standard VAE is employed to statistically extract latent space hidden in sampled data, and this latent space helps make robots controllable in feasible computational time and cost. To improve the usefulness of the latent space, this paper focuses on disentangled representation learning, e.g., $\beta$-VAE, which is the baseline for it. Starting from a Tsallis statistics perspective… CONTINUE READING
    1 Citations

    References

    SHOWING 1-10 OF 39 REFERENCES
    beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework
    • 1,454
    • Highly Influential
    Isolating Sources of Disentanglement in Variational Autoencoders
    • 391
    • PDF
    Understanding disentangling in β-VAE
    • 202
    Understanding disentangling in $\beta$-VAE
    • 184
    • PDF
    Variational Deep Embedding with Regularized Student-t Mixture Model
    • 1
    Student-t Variational Autoencoder for Robust Density Estimation
    • 9
    • PDF
    A Disentangled Recognition and Nonlinear Dynamics Model for Unsupervised Learning
    • 123
    • PDF
    Towards a Definition of Disentangled Representations
    • 106
    • PDF
    Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Images
    • 442
    • PDF