The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Label bising dalam data latihan boleh menjejaskan prestasi rangkaian saraf dalam (DNN) dengan ketara. Penyelidikan terkini tentang pembelajaran dengan label bising menggunakan sifat DNN yang dipanggil kesan hafalan untuk membahagikan data latihan kepada satu set data dengan label yang boleh dipercayai dan satu set data dengan label yang tidak boleh dipercayai. Kaedah yang memperkenalkan strategi pembelajaran separa penyeliaan membuang label yang tidak boleh dipercayai dan menetapkan pseudo-label yang dijana daripada ramalan yakin model. Setakat ini, strategi separa penyeliaan ini telah membuahkan hasil yang terbaik dalam bidang ini. Walau bagaimanapun, kami mendapati bahawa walaupun model dilatih pada data seimbang, pengedaran pseudo-label masih boleh menunjukkan ketidakseimbangan yang didorong oleh persamaan data. Selain itu, bias data dilihat berpunca daripada pembahagian data latihan menggunakan kaedah separa penyeliaan. Jika kita menangani kedua-dua jenis bias yang timbul daripada pseudo-label, kita boleh mengelakkan penurunan dalam prestasi generalisasi yang disebabkan oleh berat sebelah. bising pseudo-label. Kami mencadangkan kaedah pembelajaran dengan label bising yang memperkenalkan pelabelan pseudo yang tidak berat sebelah berdasarkan inferens sebab. Kaedah yang dicadangkan mencapai keuntungan ketepatan yang ketara dalam eksperimen pada kadar hingar yang tinggi pada penanda aras standard CIFAR-10 dan CIFAR-100.
Ryota HIGASHIMOTO
Kansai University
Soh YOSHIDA
Kansai University
Takashi HORIHATA
Kansai University
Mitsuji MUNEYASU
Kansai University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Salinan
Ryota HIGASHIMOTO, Soh YOSHIDA, Takashi HORIHATA, Mitsuji MUNEYASU, "Unbiased Pseudo-Labeling for Learning with Noisy Labels" in IEICE TRANSACTIONS on Information,
vol. E107-D, no. 1, pp. 44-48, January 2024, doi: 10.1587/transinf.2023MUL0002.
Abstract: Noisy labels in training data can significantly harm the performance of deep neural networks (DNNs). Recent research on learning with noisy labels uses a property of DNNs called the memorization effect to divide the training data into a set of data with reliable labels and a set of data with unreliable labels. Methods introducing semi-supervised learning strategies discard the unreliable labels and assign pseudo-labels generated from the confident predictions of the model. So far, this semi-supervised strategy has yielded the best results in this field. However, we observe that even when models are trained on balanced data, the distribution of the pseudo-labels can still exhibit an imbalance that is driven by data similarity. Additionally, a data bias is seen that originates from the division of the training data using the semi-supervised method. If we address both types of bias that arise from pseudo-labels, we can avoid the decrease in generalization performance caused by biased noisy pseudo-labels. We propose a learning method with noisy labels that introduces unbiased pseudo-labeling based on causal inference. The proposed method achieves significant accuracy gains in experiments at high noise rates on the standard benchmarks CIFAR-10 and CIFAR-100.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2023MUL0002/_p
Salinan
@ARTICLE{e107-d_1_44,
author={Ryota HIGASHIMOTO, Soh YOSHIDA, Takashi HORIHATA, Mitsuji MUNEYASU, },
journal={IEICE TRANSACTIONS on Information},
title={Unbiased Pseudo-Labeling for Learning with Noisy Labels},
year={2024},
volume={E107-D},
number={1},
pages={44-48},
abstract={Noisy labels in training data can significantly harm the performance of deep neural networks (DNNs). Recent research on learning with noisy labels uses a property of DNNs called the memorization effect to divide the training data into a set of data with reliable labels and a set of data with unreliable labels. Methods introducing semi-supervised learning strategies discard the unreliable labels and assign pseudo-labels generated from the confident predictions of the model. So far, this semi-supervised strategy has yielded the best results in this field. However, we observe that even when models are trained on balanced data, the distribution of the pseudo-labels can still exhibit an imbalance that is driven by data similarity. Additionally, a data bias is seen that originates from the division of the training data using the semi-supervised method. If we address both types of bias that arise from pseudo-labels, we can avoid the decrease in generalization performance caused by biased noisy pseudo-labels. We propose a learning method with noisy labels that introduces unbiased pseudo-labeling based on causal inference. The proposed method achieves significant accuracy gains in experiments at high noise rates on the standard benchmarks CIFAR-10 and CIFAR-100.},
keywords={},
doi={10.1587/transinf.2023MUL0002},
ISSN={1745-1361},
month={January},}
Salinan
TY - JOUR
TI - Unbiased Pseudo-Labeling for Learning with Noisy Labels
T2 - IEICE TRANSACTIONS on Information
SP - 44
EP - 48
AU - Ryota HIGASHIMOTO
AU - Soh YOSHIDA
AU - Takashi HORIHATA
AU - Mitsuji MUNEYASU
PY - 2024
DO - 10.1587/transinf.2023MUL0002
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E107-D
IS - 1
JA - IEICE TRANSACTIONS on Information
Y1 - January 2024
AB - Noisy labels in training data can significantly harm the performance of deep neural networks (DNNs). Recent research on learning with noisy labels uses a property of DNNs called the memorization effect to divide the training data into a set of data with reliable labels and a set of data with unreliable labels. Methods introducing semi-supervised learning strategies discard the unreliable labels and assign pseudo-labels generated from the confident predictions of the model. So far, this semi-supervised strategy has yielded the best results in this field. However, we observe that even when models are trained on balanced data, the distribution of the pseudo-labels can still exhibit an imbalance that is driven by data similarity. Additionally, a data bias is seen that originates from the division of the training data using the semi-supervised method. If we address both types of bias that arise from pseudo-labels, we can avoid the decrease in generalization performance caused by biased noisy pseudo-labels. We propose a learning method with noisy labels that introduces unbiased pseudo-labeling based on causal inference. The proposed method achieves significant accuracy gains in experiments at high noise rates on the standard benchmarks CIFAR-10 and CIFAR-100.
ER -