Published on Sat Apr 10 2021

Latent Code-Based Fusion: A Volterra Neural Network Approach

Sally Ghanem, Siddharth Roheda, Hamid Krim
0
0
0
Abstract

We propose a deep structure encoder using the recently introduced Volterra Neural Networks (VNNs) to seek a latent representation of multi-modal data whose features are jointly captured by a union of subspaces. The so-called self-representation embedding of the latent codes leads to a simplified fusion which is driven by a similarly constructed decoding. The Volterra Filter architecture achieved reduction in parameter complexity is primarily due to controlled non-linearities being introduced by the higher-order convolutions in contrast to generalized activation functions. Experimental results on two different datasets have shown a significant improvement in the clustering performance for VNNs auto-encoder over conventional Convolutional Neural Networks (CNNs) auto-encoder. In addition, we also show that the proposed approach demonstrates a much-improved sample complexity over CNN-based auto-encoder with a superb robust classification performance.

Tue Apr 17 2018
Machine Learning
Deep Multimodal Subspace Clustering Networks
We present convolutional neural network (CNN) based approaches for unsupervised multimodal subspace clustering. The proposed framework consists of three main stages. The self-expressive layer is responsible for enforcing the self-expressiveness property and acquiring an affinity matrix corresponding to data points.
0
0
0
Fri Sep 08 2017
Computer Vision
Deep Subspace Clustering Networks
We present a novel deep neural network architecture for unsupervised subspace clustering. This architecture is built upon deep auto-encoders, which are non-linearly map the input data into a latent space. We propose a novel self-expressive layer between the encoder and
0
0
0
Mon May 28 2018
Neural Networks
Deep Discriminative Latent Space for Clustering
Clustering is one of the most fundamental tasks in data analysis and machine learning. We propose to optimize an auto-encoder with respect to a discriminative pairwise loss function. We demonstrate the high accuracy obtained by the proposed method.
0
0
0
Tue Apr 30 2019
Machine Learning
Deep Spectral Clustering using Dual Autoencoder Network
Deep clustering combines embedding and clustering together to obtain optimal embedding subspace for clustering. The method can be more effective compared with conventional clustering methods.
0
0
0
Wed Apr 03 2019
Machine Learning
Multimodal Representation Learning using Deep Multiset Canonical Correlation
Deep Multiset Canonical Correlation Analysis (dMCCA) is an extension to representation learning using CCA. We use deep learning framework to learn non-linear transformations from different modalities. Unlike linear discriminant analysis, we do not need class information to learn these representations.
0
0
0
Mon Apr 22 2019
Computer Vision
Local Deep-Feature Alignment for Unsupervised Dimension Reduction
This paper presents an unsupervised deep-learning framework named Local Deep-Feature Alignment (LDFA) for dimension reduction. The advantage of the LDFA method is that it learns both local and global characteristics of the data sample set.
0
0
0