Learning to Decompose and Disentangle Representations for Video Prediction

Abstract

Our goal is to predict future video frames given a sequence of input frames. Despite large amounts of video data, this remains a challenging task because of the high-dimensionality of video frames. We address this challenge by proposing the Decompositional Disentangled Predictive Auto-Encoder (DDPAE), a framework that combines structured probabilistic models and deep networks to automatically (i) decompose the high-dimensional video that we aim to predict into components, and (ii) disentangle each component to have low-dimensional temporal dynamics that are easier to predict. Crucially, with an appropriately specified generative model of video frames, our DDPAE is able to learn both the latent decomposition and disentanglement without explicit supervision. For the Moving MNIST dataset, we show that DDPAE is able to recover the underlying components (individual digits) and disentanglement (appearance and location) as we would intuitively do. We further demonstrate that DDPAE can be applied to the Bouncing Balls dataset involving complex interactions between multiple objects to predict the video frame directly from the pixels and recover physical states without explicit supervision.

References in corpus (20)

• Auto-Encoding Variational Bayes
• Sequence to Sequence Learning with Neural Networks
• Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting
• DRAW: A Recurrent Neural Network For Image Generation
• Generating Videos with Scene Dynamics
• Action-Conditional Video Prediction using Deep Networks in Atari Games
• Decomposing Motion and Content for Natural Video Sequence Prediction
• Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning
• Video (language) modeling: a baseline for generative models of natural videos
• Unsupervised Learning for Physical Interaction through Video Prediction
• Attend, Infer, Repeat: Fast Scene Understanding with Generative Models
• Spatio-temporal video autoencoder with differentiable memory
• Unsupervised Learning of Disentangled Representations from Video
• Learning to Generate Long-term Future via Hierarchical Prediction
• A Compositional Object-Based Approach to Learning Physical Dynamics
• Visual Dynamics: Probabilistic Future Frame Synthesis via Cross Convolutional Networks
• Relational Neural Expectation Maximization: Unsupervised Discovery of Objects and their Interactions
• Transformation-Based Models of Video Sequences
• Object-Centric Representation Learning from Unlabeled Videos
• Hybrid VAE: Improving Deep Generative Models using Partial Observations

Cited by in corpus (6)

• A Review on Deep Learning Techniques for Video Prediction
• Boosting RGB-D Saliency Detection by Leveraging Unlabeled RGB Images
• Physics-as-Inverse-Graphics: Unsupervised Physical Parameter Estimation from Video
• TransMoMo: Invariance-Driven Unsupervised Video Motion Retargeting
• Motion Segmentation using Frequency Domain Transformer Networks
• Motion Selective Prediction for Video Frame Synthesis