Stacked Capsule Autoencoders

Published in

Neural Information Processing Systems(2019)

External Links:

Generate Graph

TL;DR

This work introduces an unsupervised capsule autoencoder (SCAE), which explicitly uses geometric relationships between parts to reason about objects, and finds that object capsule presences are highly informative of the object class, which leads to state-of-the-art results for un supervised classification on SVHN and MNIST.

Abstract

Objects are composed of a set of geometrically organized parts. We introduce an unsupervised capsule autoencoder (SCAE), which explicitly uses geometric relationships between parts to reason about objects. Since these relationships do not depend on the viewpoint, our model is robust to viewpoint changes. SCAE consists of two stages. In the first stage, the model predicts presences and poses of part templates directly from the image and tries to reconstruct the image by appropriately arranging the templates. In the second stage, SCAE predicts parameters of a few object capsules, which are then used to reconstruct part poses. Inference in this model is amortized and performed by off-the-shelf neural encoders, unlike in previous capsule networks. We find that object capsule presences are highly informative of the object class, which leads to state-of-the-art results for unsupervised classification on SVHN (55%) and MNIST (98.7%). The code is available at this https URL

Authors

Geoffrey E. Hinton

13 papers

Y. Teh

7 papers

S. Sabour

5 papers

References37 items

GENESIS: Generative Scene Inference and Sampling with Object-Centric Latent Representations

Multi-Object Representation Learning with Iterative Variational Inference

MONet: Unsupervised Scene Decomposition and Representation

Associative Deep Clustering: Training a Classification Network with No Labels

Set Transformer

Stacked Capsule Autoencoders

Published in

Neural Information Processing Systems(2019)

External Links:

Generate Graph

TL;DR

Abstract

Authors

Geoffrey E. Hinton

13 papers

Y. Teh

7 papers

S. Sabour

5 papers

References37 items

GENESIS: Generative Scene Inference and Sampling with Object-Centric Latent Representations

Multi-Object Representation Learning with Iterative Variational Inference

MONet: Unsupervised Scene Decomposition and Representation

Associative Deep Clustering: Training a Classification Network with No Labels

Set Transformer

Adam R. Kosiorek

2 papers

Learning deep representations by mutual information estimation and maximization

Neural Network Encapsulation

Invariant Information Distillation for Unsupervised Image Segmentation and Clustering

Group Equivariant Capsule Networks

Sequential Attend, Infer, Repeat: Generative Modelling of Moving Objects

Fast dynamic routing based on weighted kernel density estimation

Sparse Unsupervised Capsules Generalize Better

CapsuleGAN: Generative Adversarial Capsule Network

Matrix capsules with EM routing

An Optimization View on Dynamic Routing Between Capsules

Dynamic Routing Between Capsules

Dynamic Steerable Blocks in Deep Residual Networks

Learning Discrete Representations via Information Maximizing Self-Augmented Training

The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables

Layer Normalization

Attend, Infer, Repeat: Fast Scene Understanding with Generative Models

Group Equivariant Convolutional Networks

Exploiting Cyclic Symmetry in Convolutional Neural Networks

Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

Semi-supervised Learning with Ladder Networks

Spatial Transformer Networks

Deep roto-translation scattering for object classification

Optimizing Neural Networks that Generate Iimages

Transforming Auto-Encoders

Noise-contrastive estimation: A new estimation principle for unnormalized statistical models

Greedy Layer-Wise Training of Deep Networks

Some Demonstrations of the Effects of Structural Descriptions in Mental Imagery

Orientation and Form

The Hungarian method for the assignment problem

Lecture 6.5—RmsProp: Divide the gradient by a running average of its recent magnitude. COURSERA: Neural Networks for Machine Learning

Visualizing Data using t-SNE

Field of Study

Computer ScienceMathematics

Journal Information

Name

ArXiv

Volume

abs/2005.00687

Venue Information

Name

Neural Information Processing Systems

Type

conference

URL

http://neurips.cc/

Alternate Names

Neural Inf Process Syst
NeurIPS
NIPS

TL;DR

Abstract

Authors

References37 items

GENESIS: Generative Scene Inference and Sampling with Object-Centric Latent Representations

Multi-Object Representation Learning with Iterative Variational Inference

MONet: Unsupervised Scene Decomposition and Representation

Associative Deep Clustering: Training a Classification Network with No Labels

Set Transformer

TL;DR

Abstract

Authors

References37 items

GENESIS: Generative Scene Inference and Sampling with Object-Centric Latent Representations

Multi-Object Representation Learning with Iterative Variational Inference

MONet: Unsupervised Scene Decomposition and Representation

Associative Deep Clustering: Training a Classification Network with No Labels

Set Transformer

Learning deep representations by mutual information estimation and maximization

Neural Network Encapsulation

Invariant Information Distillation for Unsupervised Image Segmentation and Clustering

Group Equivariant Capsule Networks

Sequential Attend, Infer, Repeat: Generative Modelling of Moving Objects

Fast dynamic routing based on weighted kernel density estimation

Sparse Unsupervised Capsules Generalize Better

CapsuleGAN: Generative Adversarial Capsule Network

Matrix capsules with EM routing

An Optimization View on Dynamic Routing Between Capsules

Dynamic Routing Between Capsules

Dynamic Steerable Blocks in Deep Residual Networks

Learning Discrete Representations via Information Maximizing Self-Augmented Training

Steerable CNNs

The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables

Layer Normalization

Attend, Infer, Repeat: Fast Scene Understanding with Generative Models

Group Equivariant Convolutional Networks

Exploiting Cyclic Symmetry in Convolutional Neural Networks

Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

Semi-supervised Learning with Ladder Networks

Spatial Transformer Networks

Deep roto-translation scattering for object classification

Optimizing Neural Networks that Generate Iimages

Transforming Auto-Encoders

Noise-contrastive estimation: A new estimation principle for unnormalized statistical models

Greedy Layer-Wise Training of Deep Networks

Some Demonstrations of the Effects of Structural Descriptions in Mental Imagery

Orientation and Form

The Hungarian method for the assignment problem

Lecture 6.5—RmsProp: Divide the gradient by a running average of its recent magnitude. COURSERA: Neural Networks for Machine Learning

Visualizing Data using t-SNE

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names