Learning to Linearize Under Uncertainty

Abstract

Training deep feature hierarchies to solve supervised learning tasks has achieved state of the art performance on many problems in computer vision. However, a principled way in which to train such hierarchies in the unsupervised setting has remained elusive. In this work we suggest a new architecture and loss for training deep feature hierarchies that linearize the transformations observed in unlabeled natural video sequences. This is done by training a generative model to predict video frames. We also address the problem of inherent uncertainty in prediction by introducing latent variables that are non-deterministic functions of the input into the network architecture.

To appear at NIPS 2015

References in corpus (1)

• Video (language) modeling: a baseline for generative models of natural videos

Cited by in corpus (17)

• A Review on Deep Learning Techniques for Video Prediction
• Discovery of Latent 3D Keypoints via End-to-end Geometric Reasoning
• DARLA: Improving Zero-Shot Transfer in Reinforcement Learning
• Are Disentangled Representations Helpful for Abstract Visual Reasoning?
• Weakly-Supervised Disentanglement Without Compromises
• Disentangling Hate in Online Memes
• Recovering missing CFD data for high-order discretizations using deep neural networks and dynamics learning
• Disentangling Factors of Variation Using Few Labels
• Are we done with object recognition? The iCub robot's perspective
• Transporter Networks: Rearranging the Visual World for Robotic Manipulation
• On the Fairness of Disentangled Representations
• Gradient-based Training of Slow Feature Analysis by Differentiable Approximate Whitening
• Learning Actionable Representations with Goal-Conditioned Policies
• Future Frame Prediction of a Video Sequence
• Towards Object Detection from Motion
• Temporal Action Localization with Variance-Aware Networks
• Pose Augmentation: Class-agnostic Object Pose Transformation for Object Recognition