Sorry, you need to enable JavaScript to visit this website.

Neural network learning (MLR-NNLR)

TOWARDS THINNER CONVOLUTIONAL NEURAL NETWORKS THROUGH GRADUALLY GLOBAL PRUNING


Deep network pruning is an effective method to reduce the storage and computation cost of deep neural networks when applying them to resource-limited devices. Among many pruning granularities, neuron level pruning will remove redundant neurons and filters in the model and result in thinner networks. In this paper, we propose a gradually global pruning scheme for neuron level pruning. In each pruning step,

Paper Details

Authors:
Zhengtao Wang, Ce Zhu, Zhiqiang Xia, Qi Guo, Yipeng Liu
Submitted On:
15 September 2017 - 1:19pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

ICIP1701

(7)

Subscribe

[1] Zhengtao Wang, Ce Zhu, Zhiqiang Xia, Qi Guo, Yipeng Liu, " TOWARDS THINNER CONVOLUTIONAL NEURAL NETWORKS THROUGH GRADUALLY GLOBAL PRUNING", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2155. Accessed: Apr. 25, 2019.
@article{2155-17,
url = {http://sigport.org/2155},
author = {Zhengtao Wang; Ce Zhu; Zhiqiang Xia; Qi Guo; Yipeng Liu },
publisher = {IEEE SigPort},
title = { TOWARDS THINNER CONVOLUTIONAL NEURAL NETWORKS THROUGH GRADUALLY GLOBAL PRUNING},
year = {2017} }
TY - EJOUR
T1 - TOWARDS THINNER CONVOLUTIONAL NEURAL NETWORKS THROUGH GRADUALLY GLOBAL PRUNING
AU - Zhengtao Wang; Ce Zhu; Zhiqiang Xia; Qi Guo; Yipeng Liu
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2155
ER -
Zhengtao Wang, Ce Zhu, Zhiqiang Xia, Qi Guo, Yipeng Liu. (2017). TOWARDS THINNER CONVOLUTIONAL NEURAL NETWORKS THROUGH GRADUALLY GLOBAL PRUNING. IEEE SigPort. http://sigport.org/2155
Zhengtao Wang, Ce Zhu, Zhiqiang Xia, Qi Guo, Yipeng Liu, 2017. TOWARDS THINNER CONVOLUTIONAL NEURAL NETWORKS THROUGH GRADUALLY GLOBAL PRUNING. Available at: http://sigport.org/2155.
Zhengtao Wang, Ce Zhu, Zhiqiang Xia, Qi Guo, Yipeng Liu. (2017). " TOWARDS THINNER CONVOLUTIONAL NEURAL NETWORKS THROUGH GRADUALLY GLOBAL PRUNING." Web.
1. Zhengtao Wang, Ce Zhu, Zhiqiang Xia, Qi Guo, Yipeng Liu. TOWARDS THINNER CONVOLUTIONAL NEURAL NETWORKS THROUGH GRADUALLY GLOBAL PRUNING [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2155

ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING

Paper Details

Authors:
Qi Guo, Ce Zhu, Zhiqiang Xia, Zhengtao Wang, Yipeng Liu
Submitted On:
15 September 2017 - 11:50am
Short Link:
Type:
Event:
Document Year:
Cite

Document Files

ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING.pdf

(8)

Subscribe

[1] Qi Guo, Ce Zhu, Zhiqiang Xia, Zhengtao Wang, Yipeng Liu , "ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2146. Accessed: Apr. 25, 2019.
@article{2146-17,
url = {http://sigport.org/2146},
author = {Qi Guo; Ce Zhu; Zhiqiang Xia; Zhengtao Wang; Yipeng Liu },
publisher = {IEEE SigPort},
title = {ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING},
year = {2017} }
TY - EJOUR
T1 - ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING
AU - Qi Guo; Ce Zhu; Zhiqiang Xia; Zhengtao Wang; Yipeng Liu
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2146
ER -
Qi Guo, Ce Zhu, Zhiqiang Xia, Zhengtao Wang, Yipeng Liu . (2017). ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING. IEEE SigPort. http://sigport.org/2146
Qi Guo, Ce Zhu, Zhiqiang Xia, Zhengtao Wang, Yipeng Liu , 2017. ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING. Available at: http://sigport.org/2146.
Qi Guo, Ce Zhu, Zhiqiang Xia, Zhengtao Wang, Yipeng Liu . (2017). "ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING." Web.
1. Qi Guo, Ce Zhu, Zhiqiang Xia, Zhengtao Wang, Yipeng Liu . ATTRIBUTE-CONTROLLED FACE PHOTO SYNTHESIS FROM SIMPLE LINE DRAWING [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2146

Search Video Action Proposal with Recurrent and Static YOLO


In this paper, we propose a new approach for searching action proposals in unconstrained videos. Our method first produces snippet action proposals by combining state-of-the-art YOLO detector (Static YOLO) and our regression based RNN detector (Recurrent YOLO). Then, these short action proposals are integrated to form final action proposals by solving two-pass dynamic programming which maximizes actioness score and temporal smoothness concurrently.

Paper Details

Authors:
Romain Vial, Hongyuan Zhu, Yonghong Tian, Shijian Lu
Submitted On:
20 September 2017 - 10:51am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

presentation.pdf

(780)

Subscribe

[1] Romain Vial, Hongyuan Zhu, Yonghong Tian, Shijian Lu, "Search Video Action Proposal with Recurrent and Static YOLO", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2140. Accessed: Apr. 25, 2019.
@article{2140-17,
url = {http://sigport.org/2140},
author = {Romain Vial; Hongyuan Zhu; Yonghong Tian; Shijian Lu },
publisher = {IEEE SigPort},
title = {Search Video Action Proposal with Recurrent and Static YOLO},
year = {2017} }
TY - EJOUR
T1 - Search Video Action Proposal with Recurrent and Static YOLO
AU - Romain Vial; Hongyuan Zhu; Yonghong Tian; Shijian Lu
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2140
ER -
Romain Vial, Hongyuan Zhu, Yonghong Tian, Shijian Lu. (2017). Search Video Action Proposal with Recurrent and Static YOLO. IEEE SigPort. http://sigport.org/2140
Romain Vial, Hongyuan Zhu, Yonghong Tian, Shijian Lu, 2017. Search Video Action Proposal with Recurrent and Static YOLO. Available at: http://sigport.org/2140.
Romain Vial, Hongyuan Zhu, Yonghong Tian, Shijian Lu. (2017). "Search Video Action Proposal with Recurrent and Static YOLO." Web.
1. Romain Vial, Hongyuan Zhu, Yonghong Tian, Shijian Lu. Search Video Action Proposal with Recurrent and Static YOLO [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2140

Foveated Neural Network: Gaze Prediction On Egocentric Videos


A novel deep convolution neural network, named as Foveated Neural Network (FNN), is proposed to predict gaze on current frames in egocentric videos. The retina-like visual inputs from the region of interest on the previous frame get analysed and encoded. The fusion of the hidden representation of the previous frame and the feature maps of the current frame guides the gaze prediction process on the current frame. In order to simulate motions, we also include the dense optical flow between these adjacent frames as additional inputs to FNN.

Paper Details

Authors:
Mengmi Zhang, Keng-Teck Ma, Joo-Hwee Lim, Qi Zhao
Submitted On:
15 September 2017 - 4:11am
Short Link:
Type:
Event:
Paper Code:
Document Year:
Cite

Document Files

ICIP17_poster.pdf

(371)

Subscribe

[1] Mengmi Zhang, Keng-Teck Ma, Joo-Hwee Lim, Qi Zhao, "Foveated Neural Network: Gaze Prediction On Egocentric Videos", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2100. Accessed: Apr. 25, 2019.
@article{2100-17,
url = {http://sigport.org/2100},
author = {Mengmi Zhang; Keng-Teck Ma; Joo-Hwee Lim; Qi Zhao },
publisher = {IEEE SigPort},
title = {Foveated Neural Network: Gaze Prediction On Egocentric Videos},
year = {2017} }
TY - EJOUR
T1 - Foveated Neural Network: Gaze Prediction On Egocentric Videos
AU - Mengmi Zhang; Keng-Teck Ma; Joo-Hwee Lim; Qi Zhao
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2100
ER -
Mengmi Zhang, Keng-Teck Ma, Joo-Hwee Lim, Qi Zhao. (2017). Foveated Neural Network: Gaze Prediction On Egocentric Videos. IEEE SigPort. http://sigport.org/2100
Mengmi Zhang, Keng-Teck Ma, Joo-Hwee Lim, Qi Zhao, 2017. Foveated Neural Network: Gaze Prediction On Egocentric Videos. Available at: http://sigport.org/2100.
Mengmi Zhang, Keng-Teck Ma, Joo-Hwee Lim, Qi Zhao. (2017). "Foveated Neural Network: Gaze Prediction On Egocentric Videos." Web.
1. Mengmi Zhang, Keng-Teck Ma, Joo-Hwee Lim, Qi Zhao. Foveated Neural Network: Gaze Prediction On Egocentric Videos [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2100

LEARNING TO GENERATE IMAGES WITH PERCEPTUAL SIMILARITY METRICS (POSTER)

Paper Details

Authors:
Jake Snell, Karl Ridgeway, Renjie Liao, Brett D. Roads, Michael C. Mozer, Richard S. Zemel
Submitted On:
14 September 2017 - 10:45pm
Short Link:
Type:
Event:
Presenter's Name:
Document Year:
Cite

Document Files

poster_2944.pdf

(172)

Subscribe

[1] Jake Snell, Karl Ridgeway, Renjie Liao, Brett D. Roads, Michael C. Mozer, Richard S. Zemel, "LEARNING TO GENERATE IMAGES WITH PERCEPTUAL SIMILARITY METRICS (POSTER)", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2073. Accessed: Apr. 25, 2019.
@article{2073-17,
url = {http://sigport.org/2073},
author = {Jake Snell; Karl Ridgeway; Renjie Liao; Brett D. Roads; Michael C. Mozer; Richard S. Zemel },
publisher = {IEEE SigPort},
title = {LEARNING TO GENERATE IMAGES WITH PERCEPTUAL SIMILARITY METRICS (POSTER)},
year = {2017} }
TY - EJOUR
T1 - LEARNING TO GENERATE IMAGES WITH PERCEPTUAL SIMILARITY METRICS (POSTER)
AU - Jake Snell; Karl Ridgeway; Renjie Liao; Brett D. Roads; Michael C. Mozer; Richard S. Zemel
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2073
ER -
Jake Snell, Karl Ridgeway, Renjie Liao, Brett D. Roads, Michael C. Mozer, Richard S. Zemel. (2017). LEARNING TO GENERATE IMAGES WITH PERCEPTUAL SIMILARITY METRICS (POSTER). IEEE SigPort. http://sigport.org/2073
Jake Snell, Karl Ridgeway, Renjie Liao, Brett D. Roads, Michael C. Mozer, Richard S. Zemel, 2017. LEARNING TO GENERATE IMAGES WITH PERCEPTUAL SIMILARITY METRICS (POSTER). Available at: http://sigport.org/2073.
Jake Snell, Karl Ridgeway, Renjie Liao, Brett D. Roads, Michael C. Mozer, Richard S. Zemel. (2017). "LEARNING TO GENERATE IMAGES WITH PERCEPTUAL SIMILARITY METRICS (POSTER)." Web.
1. Jake Snell, Karl Ridgeway, Renjie Liao, Brett D. Roads, Michael C. Mozer, Richard S. Zemel. LEARNING TO GENERATE IMAGES WITH PERCEPTUAL SIMILARITY METRICS (POSTER) [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2073

PERSON RE-IDENTIFICATION USING VISUAL ATTENTION


Despite recent attempts for solving the person re-identification problem, it remains a challenging task since a person’s appearance can vary significantly when large variations in view angle, human pose and illumination are involved. The concept of attention is one of the most interesting recent architectural innovations in neural networks. Inspired by that, in this paper we propose a novel approach based on using a gradient-based attention mechanism in deep convolution neural network for solving the person re-identification problem.

Paper Details

Authors:
Hairong Qi
Submitted On:
14 September 2017 - 4:12pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Poster

(173)

Subscribe

[1] Hairong Qi, "PERSON RE-IDENTIFICATION USING VISUAL ATTENTION", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2046. Accessed: Apr. 25, 2019.
@article{2046-17,
url = {http://sigport.org/2046},
author = {Hairong Qi },
publisher = {IEEE SigPort},
title = {PERSON RE-IDENTIFICATION USING VISUAL ATTENTION},
year = {2017} }
TY - EJOUR
T1 - PERSON RE-IDENTIFICATION USING VISUAL ATTENTION
AU - Hairong Qi
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2046
ER -
Hairong Qi. (2017). PERSON RE-IDENTIFICATION USING VISUAL ATTENTION. IEEE SigPort. http://sigport.org/2046
Hairong Qi, 2017. PERSON RE-IDENTIFICATION USING VISUAL ATTENTION. Available at: http://sigport.org/2046.
Hairong Qi. (2017). "PERSON RE-IDENTIFICATION USING VISUAL ATTENTION." Web.
1. Hairong Qi. PERSON RE-IDENTIFICATION USING VISUAL ATTENTION [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2046

MOTION BLUR REMOVAL VIA COUPLED AUTOENCODER

Paper Details

Authors:
Brojeshwar Bhwomick , Angshul Majumdar
Submitted On:
14 September 2017 - 7:04am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

ICIP_ppt.pdf

(186)

Subscribe

[1] Brojeshwar Bhwomick , Angshul Majumdar, "MOTION BLUR REMOVAL VIA COUPLED AUTOENCODER", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2014. Accessed: Apr. 25, 2019.
@article{2014-17,
url = {http://sigport.org/2014},
author = {Brojeshwar Bhwomick ; Angshul Majumdar },
publisher = {IEEE SigPort},
title = {MOTION BLUR REMOVAL VIA COUPLED AUTOENCODER},
year = {2017} }
TY - EJOUR
T1 - MOTION BLUR REMOVAL VIA COUPLED AUTOENCODER
AU - Brojeshwar Bhwomick ; Angshul Majumdar
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2014
ER -
Brojeshwar Bhwomick , Angshul Majumdar. (2017). MOTION BLUR REMOVAL VIA COUPLED AUTOENCODER. IEEE SigPort. http://sigport.org/2014
Brojeshwar Bhwomick , Angshul Majumdar, 2017. MOTION BLUR REMOVAL VIA COUPLED AUTOENCODER. Available at: http://sigport.org/2014.
Brojeshwar Bhwomick , Angshul Majumdar. (2017). "MOTION BLUR REMOVAL VIA COUPLED AUTOENCODER." Web.
1. Brojeshwar Bhwomick , Angshul Majumdar. MOTION BLUR REMOVAL VIA COUPLED AUTOENCODER [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2014

LEARNING AUTOENCODERS WITH LOW-RANK WEIGHTS

Paper Details

Authors:
Angshul Majumdar
Submitted On:
14 September 2017 - 7:00am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Poster for ICIP Kavya_1x2M_120917_Prepress.pdf

(7)

Subscribe

[1] Angshul Majumdar, "LEARNING AUTOENCODERS WITH LOW-RANK WEIGHTS", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2013. Accessed: Apr. 25, 2019.
@article{2013-17,
url = {http://sigport.org/2013},
author = {Angshul Majumdar },
publisher = {IEEE SigPort},
title = {LEARNING AUTOENCODERS WITH LOW-RANK WEIGHTS},
year = {2017} }
TY - EJOUR
T1 - LEARNING AUTOENCODERS WITH LOW-RANK WEIGHTS
AU - Angshul Majumdar
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2013
ER -
Angshul Majumdar. (2017). LEARNING AUTOENCODERS WITH LOW-RANK WEIGHTS. IEEE SigPort. http://sigport.org/2013
Angshul Majumdar, 2017. LEARNING AUTOENCODERS WITH LOW-RANK WEIGHTS. Available at: http://sigport.org/2013.
Angshul Majumdar. (2017). "LEARNING AUTOENCODERS WITH LOW-RANK WEIGHTS." Web.
1. Angshul Majumdar. LEARNING AUTOENCODERS WITH LOW-RANK WEIGHTS [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2013

UNDERSTANDING NEURAL-NETWORK DENOISERS THROUGH AN ACTIVATION FUNCTION PERSPECTIVE

Paper Details

Authors:
Yuxiang Li, Bo Zhang, Raoul Florent
Submitted On:
14 September 2017 - 6:38am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

poster_icip2017a.pdf

(171)

Subscribe

[1] Yuxiang Li, Bo Zhang, Raoul Florent, "UNDERSTANDING NEURAL-NETWORK DENOISERS THROUGH AN ACTIVATION FUNCTION PERSPECTIVE", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2009. Accessed: Apr. 25, 2019.
@article{2009-17,
url = {http://sigport.org/2009},
author = {Yuxiang Li; Bo Zhang; Raoul Florent },
publisher = {IEEE SigPort},
title = {UNDERSTANDING NEURAL-NETWORK DENOISERS THROUGH AN ACTIVATION FUNCTION PERSPECTIVE},
year = {2017} }
TY - EJOUR
T1 - UNDERSTANDING NEURAL-NETWORK DENOISERS THROUGH AN ACTIVATION FUNCTION PERSPECTIVE
AU - Yuxiang Li; Bo Zhang; Raoul Florent
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2009
ER -
Yuxiang Li, Bo Zhang, Raoul Florent. (2017). UNDERSTANDING NEURAL-NETWORK DENOISERS THROUGH AN ACTIVATION FUNCTION PERSPECTIVE. IEEE SigPort. http://sigport.org/2009
Yuxiang Li, Bo Zhang, Raoul Florent, 2017. UNDERSTANDING NEURAL-NETWORK DENOISERS THROUGH AN ACTIVATION FUNCTION PERSPECTIVE. Available at: http://sigport.org/2009.
Yuxiang Li, Bo Zhang, Raoul Florent. (2017). "UNDERSTANDING NEURAL-NETWORK DENOISERS THROUGH AN ACTIVATION FUNCTION PERSPECTIVE." Web.
1. Yuxiang Li, Bo Zhang, Raoul Florent. UNDERSTANDING NEURAL-NETWORK DENOISERS THROUGH AN ACTIVATION FUNCTION PERSPECTIVE [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2009

Multi Layer Multi Objective Extreme Learning Machine


Fully connected multi layer neural networks such as Deep Boltzmann Machines (DBM) performs better than fully connected single layer neural networks in image classification tasks and has a smaller number of hidden layer neurons than Extreme Learning Machine (ELM) based fully connected multi layer neural networks such as Multi Layer ELM (ML-ELM) and Hierarchical ELM (H-ELM) However, ML-ELM and H-ELM has a smaller training time than DBM.

Paper Details

Authors:
Chamara Kasun Liyanaarachchi Lekamalage, Kang Song, Guang-Bin Huang, Dongshun Cui and Ken Liang
Submitted On:
12 September 2017 - 11:22pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

MLMO_ELM_presentation.pdf

(200)

Subscribe

[1] Chamara Kasun Liyanaarachchi Lekamalage, Kang Song, Guang-Bin Huang, Dongshun Cui and Ken Liang, "Multi Layer Multi Objective Extreme Learning Machine", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/1957. Accessed: Apr. 25, 2019.
@article{1957-17,
url = {http://sigport.org/1957},
author = {Chamara Kasun Liyanaarachchi Lekamalage; Kang Song; Guang-Bin Huang; Dongshun Cui and Ken Liang },
publisher = {IEEE SigPort},
title = {Multi Layer Multi Objective Extreme Learning Machine},
year = {2017} }
TY - EJOUR
T1 - Multi Layer Multi Objective Extreme Learning Machine
AU - Chamara Kasun Liyanaarachchi Lekamalage; Kang Song; Guang-Bin Huang; Dongshun Cui and Ken Liang
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/1957
ER -
Chamara Kasun Liyanaarachchi Lekamalage, Kang Song, Guang-Bin Huang, Dongshun Cui and Ken Liang. (2017). Multi Layer Multi Objective Extreme Learning Machine. IEEE SigPort. http://sigport.org/1957
Chamara Kasun Liyanaarachchi Lekamalage, Kang Song, Guang-Bin Huang, Dongshun Cui and Ken Liang, 2017. Multi Layer Multi Objective Extreme Learning Machine. Available at: http://sigport.org/1957.
Chamara Kasun Liyanaarachchi Lekamalage, Kang Song, Guang-Bin Huang, Dongshun Cui and Ken Liang. (2017). "Multi Layer Multi Objective Extreme Learning Machine." Web.
1. Chamara Kasun Liyanaarachchi Lekamalage, Kang Song, Guang-Bin Huang, Dongshun Cui and Ken Liang. Multi Layer Multi Objective Extreme Learning Machine [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/1957

Pages