Understanding Visual Concepts with Continuation Learning

Abstract

We introduce a neural network architecture and a learning algorithm to produce factorized symbolic representations. We propose to learn these concepts by observing consecutive frames, letting all the components of the hidden representation except a small discrete set (gating units) be predicted from the previous frame, and let the factors of variation in the next frame be represented entirely by these discrete gated units (corresponding to symbolic representations). We demonstrate the efficacy of our approach on datasets of faces undergoing 3D transformations and Atari 2600 games.

Under review as a workshop paper for ICLR 2016

References in corpus (4)

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Deep Convolutional Inverse Graphics Network
• Disentangled Representations in Neural Models

Cited by in corpus (20)

• Hierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation
• Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning
• Flexible Neural Representation for Physics Prediction
• DARLA: Improving Zero-Shot Transfer in Reinforcement Learning
• Early Visual Concept Learning with Unsupervised Deep Learning
• Are Disentangled Representations Helpful for Abstract Visual Reasoning?
• Unsupervised Learning of Object Keypoints for Perception and Control
• Weakly-Supervised Disentanglement Without Compromises
• Deep Successor Reinforcement Learning
• Learning Disentangled Joint Continuous and Discrete Representations
• Variational Inference of Disentangled Latent Concepts from Unlabeled Observations
• Time-Contrastive Networks: Self-Supervised Learning from Video
• Disentangling Factors of Variation Using Few Labels
• Entity Abstraction in Visual Model-Based Reinforcement Learning
• Relevance Factor VAE: Learning and Identifying Disentangled Factors
• Disentangled and Controllable Face Image Generation via 3D Imitative-Contrastive Learning
• On the Fairness of Disentangled Representations
• TransMoMo: Invariance-Driven Unsupervised Video Motion Retargeting
• Odd-One-Out Representation Learning
• Disentangling Video with Independent Prediction