How can deep rectifier networks achieve linear separability and preserve distances?

QR Code

How can deep rectifier networks achieve linear separability and preserve distances?

This paper investigates how hidden layers of deep rectifier networks are capable of transforming two or more pattern sets to be linearly separable while preserving the distances with a guaranteed degree, and proves the universal classification power of such distance preserving rectifier networks. Th...

Full description

Bibliographic Details
Main Authors:	An, Senjian, Boussaid, F., Bennamoun, M.
Format:	Conference Paper
Published:	2015
Online Access:	http://hdl.handle.net/20.500.11937/69676

Similar Items

Exploiting layerwise convexity of rectifier networks with sign constrained weights
by: An, Senjian, et al.
Published: (2018)

Sign constrained rectifier networks with applications to pattern decompositions
by: An, Senjian, et al.
Published: (2015)

Contractive rectifier networks for nonlinear maximum margin classification
by: An, Senjian, et al.
Published: (2015)

Human interaction prediction using deep temporal features
by: Ke, Q., et al.
Published: (2016)

SkeletonNet: Mining Deep Part Features for 3-D Action Recognition
by: Ke, Q., et al.
Published: (2017)

Deep Learning for Coral Classification
by: Mahmood, A., et al.
Published: (2017)

Classification of corals in reflectance and fluorescence images using convolutional neural network representations
by: Xu, L., et al.
Published: (2018)

Deep Image Representations for Coral Image Classification
by: Mahmood, A., et al.
Published: (2018)

Automatic annotation of coral reefs using deep learning
by: Mahmood, A., et al.
Published: (2016)

Quantitative error analysis of bilateral filtering
by: An, Senjian, et al.
Published: (2015)

Deep reconstruction models for image set classification
by: Hayat, M., et al.
Published: (2015)

Learning non-linear reconstruction models for image set classification
by: Hayat, M., et al.
Published: (2014)

A new representation of skeleton sequences for 3D action recognition
by: Ke, Q., et al.
Published: (2017)

Leveraging Structural Context Models and Ranking Score Fusion for Human Interaction Prediction
by: Ke, Q., et al.
Published: (2018)

Learning Clip Representations for Skeleton-Based 3D Action Recognition
by: Ke, Q., et al.
Published: (2018)

Preservers of pairs of bivectors with bounded distance
by: Lim, Ming-Huat, et al.
Published: (2009)

Preservers of matrix pairs with bounded distance
by: Lim, Ming-Huat, et al.
Published: (2007)

Identity Adaptation for Person Re-Identification
by: Ke, Q., et al.
Published: (2018)

Computer vision for human-machine interaction
by: Ke, Q., et al.
Published: (2018)

Resfeats: Residual network based features for image classification
by: Mahmood, A., et al.
Published: (2018)

A review of effective doctoral supervision: what is it and how can we achieve it?
by: Masek, Alias, et al.
Published: (2020)

Minimising mental stress among workers: how can good corporate citizens achieve this?
by: Ciccarelli, Marina, et al.
Published: (2009)

Z-source matrix rectifier
by: Hongchen, Liu, et al.
Published: (2016)

Exploiting side information in locality preserving projection
by: An, Senjian, et al.
Published: (2008)

Separation can be a good thing
by: Menon, Sandhya
Published: (2020)

Deep residual network framework for structural health monitoring
by: Wang, Ruhua, et al.
Published: (2021)

Students of low academic achievement – Their personality, mental abilities and academic performance: How counsellor can help?
by: Dzulkifli, Mariam Adawiah, et al.
Published: (2012)

Coral classification with hybrid feature representations
by: Mahmood, A., et al.
Published: (2016)

Three phase boost rectifier design
by: Saher, Rohaizan
Published: (2012)

How can job-sharing in the NHS be used by women as an effective practice to help them achieve their career ambitions?
by: Morgan, Alistair
Published: (2018)

An efficient semi-sigmoidal non-linear activation function approach for deep neural networks
by: Chieng, Hock Hung
Published: (2022)

How effective is in vitro fertilization, and how can it be improved?
by: Stewart, L., et al.
Published: (2011)

How elastic a little can be and how much a little can do in Chinese
by: Zhang, Grace
Published: (2016)

How IT can enhance professional practice with regards to smart-phones and social networks?
by: Hassan, Nuradin Ibrahim, et al.
Published: (2016)

Abdullah : Looking back helps rectify weaknesses
Published: (2008)

Abdullah : Looking back helps rectify weaknesses
Published: (2009)

The three phase rectifier with harmonic injection current
by: Misron, Norhisam, et al.
Published: (2010)

Dynamic and control analysis of modular multi-parallel rectifiers (MMR)
by: Zare, Firuz, et al.
Published: (2017)

Examining How Takaful Operators Can Achieve Business Sustainability through Micro-Takaful from A Malaysian Perspective
by: Khalid, Khaliza
Published: (2013)

Reverse training: An efficient approach for image set classification
by: Hayat, M., et al.
Published: (2014)

Cannot write session to /tmp/vufind_sessions/sess_esdlk2j8fbah6tcqcr0fidrh0k