Abstract:Capsule networks (CapsNets) aim to parse images into a hierarchy of objects, parts, and their relations using a two-step process involving part-whole transformation and hierarchical component routing. However, this hierarchical relationship modeling is computationally expensive, which has limited the wider use of CapsNet despite its potential advantages. The current state of CapsNet models primarily focuses on comparing their performance with capsule baselines, falling short of achieving the same level of proficiency as deep CNN variants in intricate tasks. To address this limitation, we present an efficient approach for learning capsules that surpasses canonical baseline models and even demonstrates superior performance compared to high-performing convolution models. Our contribution can be outlined in two aspects: firstly, we introduce a group of subcapsules onto which an input vector is projected. Subsequently, we present the Hybrid Gromov-Wasserstein framework, which initially quantifies the dissimilarity between the input and the components modeled by the subcapsules, followed by determining their alignment degree through optimal transport. This innovative mechanism capitalizes on new insights into defining alignment between the input and subcapsules, based on the similarity of their respective component distributions. This approach enhances CapsNets' capacity to learn from intricate, high-dimensional data while retaining their interpretability and hierarchical structure. Our proposed model offers two distinct advantages: (i) its lightweight nature facilitates the application of capsules to more intricate vision tasks, including object detection; (ii) it outperforms baseline approaches in these demanding tasks.

Boosting the Robustness of Capsule Networks with Diverse Ensemble

AdvCapsNet: to Defense Adversarial Attacks Based on Capsule Networks

DE-CapsNet: A Diverse Enhanced Capsule Network with Disperse Dynamic Routing

Improving Adversarial Robustness Via Promoting Ensemble Diversity

Improving Adversarial Robustness via Promoting Ensemble Diversity.

DVERGE: Diversifying Vulnerabilities for Enhanced Robust Generation of Ensembles

Exploring Model Learning Heterogeneity for Boosting Ensemble Robustness

RobCaps: Evaluating the Robustness of Capsule Networks against Affine Transformations and Adversarial Attacks

Improving the Robustness of Capsule Networks to Image Affine Transformations

Robust triple extraction with cascade bidirectional capsule network

Capsule Network is Not More Robust than Convolutional Network

Increasing the adversarial robustness and explainability of capsule networks with $γ$-capsules

A Multi-prototype Capsule Network for Image Recognition with High Intra-class Variations

Dynamic ensemble selection based on Deep Neural Network Uncertainty Estimation for Adversarial Robustness

Improving Model Robustness Against Adversarial Examples with Redundant Fully Connected Layer.

Hybrid Gromov-Wasserstein Embedding for Capsule Learning

Pushing the Limits of Capsule Networks

On the Certified Robustness for Ensemble Models and Beyond

Diversity supporting robustness: Enhancing adversarial robustness via differentiated ensemble predictions

Capsule Network Performance on Complex Data

Adaptive Capsule Network