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

facebooktwittermailshare

CYCLIC ANNEALING TRAINING CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION WITH NOISY LABELS

Abstract: 

Noisy labels modeling makes a convolutional neural network (CNN) more robust for the image classification problem. However, current noisy labels modeling methods usually require an expectation-maximization (EM) based procedure to optimize the parameters, which is computationally expensive. In this paper, we utilize a fast annealing training method to speed up the CNN training in every M-step. Since the training is repeated executed along the entire EM optimization path and obtain many local minimal CNN models from every training cycle, we name it as the Cyclic Annealing Training (CAT) approach. In addition to reducing the training time, CAT can further bagging all the local minimal CNN models at the test time to improve the performance of classification. We evaluate the proposed method on several image classification datasets with different noisy labels patterns, and the results show that our CAT approach outperforms state-of-the-art noisy labels modeling methods.

up
0 users have voted:

Paper Details

Authors:
Jiawei Li, Tao Dai, Qingtao Tang, Yeli Xing, Shu-Tao Xia
Submitted On:
8 October 2018 - 5:30am
Short Link:
Type:
Presentation Slides
Event:
Presenter's Name:
Jiawei Li
Paper Code:
2605, MA.L1.5
Document Year:
2018
Cite

Document Files

MA.L1.5_2605_CAT_CNN_NL_v4.pdf

(72)

Subscribe

[1] Jiawei Li, Tao Dai, Qingtao Tang, Yeli Xing, Shu-Tao Xia, "CYCLIC ANNEALING TRAINING CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION WITH NOISY LABELS", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3506. Accessed: Aug. 17, 2019.
@article{3506-18,
url = {http://sigport.org/3506},
author = {Jiawei Li; Tao Dai; Qingtao Tang; Yeli Xing; Shu-Tao Xia },
publisher = {IEEE SigPort},
title = {CYCLIC ANNEALING TRAINING CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION WITH NOISY LABELS},
year = {2018} }
TY - EJOUR
T1 - CYCLIC ANNEALING TRAINING CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION WITH NOISY LABELS
AU - Jiawei Li; Tao Dai; Qingtao Tang; Yeli Xing; Shu-Tao Xia
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3506
ER -
Jiawei Li, Tao Dai, Qingtao Tang, Yeli Xing, Shu-Tao Xia. (2018). CYCLIC ANNEALING TRAINING CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION WITH NOISY LABELS. IEEE SigPort. http://sigport.org/3506
Jiawei Li, Tao Dai, Qingtao Tang, Yeli Xing, Shu-Tao Xia, 2018. CYCLIC ANNEALING TRAINING CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION WITH NOISY LABELS. Available at: http://sigport.org/3506.
Jiawei Li, Tao Dai, Qingtao Tang, Yeli Xing, Shu-Tao Xia. (2018). "CYCLIC ANNEALING TRAINING CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION WITH NOISY LABELS." Web.
1. Jiawei Li, Tao Dai, Qingtao Tang, Yeli Xing, Shu-Tao Xia. CYCLIC ANNEALING TRAINING CONVOLUTIONAL NEURAL NETWORKS FOR IMAGE CLASSIFICATION WITH NOISY LABELS [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3506