Corpus ID: 67855765

Video Extrapolation with an Invertible Linear Embedding

@article{Pottorff2019VideoEW,
  title={Video Extrapolation with an Invertible Linear Embedding},
  author={Robert Pottorff and Jared Nielsen and D. Wingate},
  journal={ArXiv},
  year={2019},
  volume={abs/1903.00133}
}
  • Robert Pottorff, Jared Nielsen, D. Wingate
  • Published 2019
  • Computer Science
  • ArXiv
  • We predict future video frames from complex dynamic scenes, using an invertible neural network as the encoder of a nonlinear dynamic system with latent linear state evolution. Our invertible linear embedding (ILE) demonstrates successful learning, prediction and latent state inference. In contrast to other approaches, ILE does not use any explicit reconstruction loss or simplistic pixel-space assumptions. Instead, it leverages invertibility to optimize the likelihood of image sequences exactly… CONTINUE READING
    2 Citations
    Video Interpolation and Prediction with Unsupervised Landmarks
    • 2
    • PDF
    A Review on Deep Learning Techniques for Video Prediction
    • 1
    • PDF

    References

    SHOWING 1-10 OF 29 REFERENCES
    Deep multi-scale video prediction beyond mean square error
    • 1,116
    • Highly Influential
    • PDF
    Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Images
    • 440
    • PDF
    NICE: Non-linear Independent Components Estimation
    • 670
    • PDF
    Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning
    • 467
    • PDF
    SDC-Net: Video Prediction Using Spatially-Displaced Convolution
    • 38
    • PDF
    Glow: Generative Flow with Invertible 1x1 Convolutions
    • 758
    • PDF
    A Dataset To Evaluate The Representations Learned By Video Prediction Models
    • 6
    • PDF
    Density estimation using Real NVP
    • 982
    • PDF