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

facebooktwittermailshare

A Scalable Convolutional Neural Network for Task-specified Scenarios via Knowledge Distillation

Abstract: 

In this paper, we explore the redundancy in convolutional neural network, which scales with the complexity of vision tasks. Considering that many front-end visual systems are interested in only a limited range of visual targets, the removing of task-specified network redundancy can promote a wide range of potential applications. We propose a task-specified knowledge distillation algorithm to derive a simplified model with pre-set computation cost and minimized accuracy loss, which suits the resource constraint front-end systems well. Experiments on the MNIST and CIFAR10 datasets demonstrate the feasibility of the proposed approach as well as the existence of task-specified redundancy.

up
0 users have voted:

Paper Details

Authors:
Mengnan Shi, Fei Qin, Qixiang Ye, Zhenjun Han, Jianbin Jiao
Submitted On:
12 March 2017 - 8:20pm
Short Link:
Type:
Poster
Event:
Presenter's Name:
Mengnan Shi
Paper Code:
MLSP-P2.6
Document Year:
2017
Cite

Document Files

ICASSP2017 poster 80cm.pdf

(52 downloads)

Subscribe

[1] Mengnan Shi, Fei Qin, Qixiang Ye, Zhenjun Han, Jianbin Jiao, "A Scalable Convolutional Neural Network for Task-specified Scenarios via Knowledge Distillation", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/1751. Accessed: Sep. 21, 2017.
@article{1751-17,
url = {http://sigport.org/1751},
author = {Mengnan Shi; Fei Qin; Qixiang Ye; Zhenjun Han; Jianbin Jiao },
publisher = {IEEE SigPort},
title = {A Scalable Convolutional Neural Network for Task-specified Scenarios via Knowledge Distillation},
year = {2017} }
TY - EJOUR
T1 - A Scalable Convolutional Neural Network for Task-specified Scenarios via Knowledge Distillation
AU - Mengnan Shi; Fei Qin; Qixiang Ye; Zhenjun Han; Jianbin Jiao
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/1751
ER -
Mengnan Shi, Fei Qin, Qixiang Ye, Zhenjun Han, Jianbin Jiao. (2017). A Scalable Convolutional Neural Network for Task-specified Scenarios via Knowledge Distillation. IEEE SigPort. http://sigport.org/1751
Mengnan Shi, Fei Qin, Qixiang Ye, Zhenjun Han, Jianbin Jiao, 2017. A Scalable Convolutional Neural Network for Task-specified Scenarios via Knowledge Distillation. Available at: http://sigport.org/1751.
Mengnan Shi, Fei Qin, Qixiang Ye, Zhenjun Han, Jianbin Jiao. (2017). "A Scalable Convolutional Neural Network for Task-specified Scenarios via Knowledge Distillation." Web.
1. Mengnan Shi, Fei Qin, Qixiang Ye, Zhenjun Han, Jianbin Jiao. A Scalable Convolutional Neural Network for Task-specified Scenarios via Knowledge Distillation [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/1751