Paper Abstract and Keywords |
Presentation |
2022-10-14 10:00
Robust Semi-Supervised Learning for Noisy Labels Using Early-learning Regularization and Weighted Loss Ryota Higashimoto, Soh Yoshida, Mitsuji Muneyasu (Kansai Univ.) |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Training Deep Neural Networks (DNNs) on datasets with incorrect labels (label noise) is an important challenge. In the presence of label noise, DNNs adapt to training samples with correct labels in the early training phase and adapt to samples with label noise in the late training phase. Recently, methods based on semi-supervised learning have shown promise in exploiting this property. In this conventional method, the sample loss distribution is modeled by a Gaussian mixture model, and the DNN is trained by assuming that samples with label noise are unlabeled. On the other hand, the accuracy of the division into labeled and unlabeled samples depends on the performance of the DNN during warm-up. In addition, the conventional method cannot consider the effect of the number of labeled samples, which varies with the percentage of label noise in the training data. In this paper, we introduce Early-learning Regularization (ELR), which suppresses the DNN's adaptation to label noise in the initial learning phase. Furthermore, we propose a learning method that introduces a weighted loss that can dynamically change the loss depending on the fraction of labeled samples. To verify the effectiveness of the proposed method, we conducted experiments using the CIFAR-100 image classification dataset with pseudo-label noise. The results show that the proposed method outperforms the existing methods on CIFAR-100 with 20textasciitilde90% label noise. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Learning with noisy labels / Semi-supervised learning / Early-learning regularization / weighted loss / / / / |
Reference Info. |
ITE Tech. Rep. |
Paper # |
|
Date of Issue |
|
ISSN |
|
Download PDF |
|
Conference Information |
Committee |
BCT IEICE-SIS |
Conference Date |
2022-10-13 - 2022-10-14 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
Hachinohe Institute of Technology |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
System Implementation Technology, Short Range Wireless Systems, Smart Multimedia Systems, Broadcasting Technology, etc. |
Paper Information |
Registration To |
IEICE-SIS |
Conference Code |
2022-10-SIS-BCT |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Robust Semi-Supervised Learning for Noisy Labels Using Early-learning Regularization and Weighted Loss |
Sub Title (in English) |
|
Keyword(1) |
Learning with noisy labels |
Keyword(2) |
Semi-supervised learning |
Keyword(3) |
Early-learning regularization |
Keyword(4) |
weighted loss |
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Ryota Higashimoto |
1st Author's Affiliation |
Kansai University (Kansai Univ.) |
2nd Author's Name |
Soh Yoshida |
2nd Author's Affiliation |
Kansai University (Kansai Univ.) |
3rd Author's Name |
Mitsuji Muneyasu |
3rd Author's Affiliation |
Kansai University (Kansai Univ.) |
4th Author's Name |
|
4th Author's Affiliation |
() |
5th Author's Name |
|
5th Author's Affiliation |
() |
6th Author's Name |
|
6th Author's Affiliation |
() |
7th Author's Name |
|
7th Author's Affiliation |
() |
8th Author's Name |
|
8th Author's Affiliation |
() |
9th Author's Name |
|
9th Author's Affiliation |
() |
10th Author's Name |
|
10th Author's Affiliation |
() |
11th Author's Name |
|
11th Author's Affiliation |
() |
12th Author's Name |
|
12th Author's Affiliation |
() |
13th Author's Name |
|
13th Author's Affiliation |
() |
14th Author's Name |
|
14th Author's Affiliation |
() |
15th Author's Name |
|
15th Author's Affiliation |
() |
16th Author's Name |
|
16th Author's Affiliation |
() |
17th Author's Name |
|
17th Author's Affiliation |
() |
18th Author's Name |
|
18th Author's Affiliation |
() |
19th Author's Name |
|
19th Author's Affiliation |
() |
20th Author's Name |
|
20th Author's Affiliation |
() |
Speaker |
Author-1 |
Date Time |
2022-10-14 10:00:00 |
Presentation Time |
20 minutes |
Registration for |
IEICE-SIS |
Paper # |
|
Volume (vol) |
vol.46 |
Number (no) |
|
Page |
|
#Pages |
|
Date of Issue |
|