Label Noise-Robust Learning using a Confidence-Based Sieving Strategy

In learning tasks with label noise, improving model robustness against overfitting is a pivotal challenge because the model eventually memorizes labels, including the noisy ones. Identifying the samples with noisy labels and preventing the model from learning them is a promising approach to address this challenge. When training with noisy labels, the per-class confidence scores of the model, represented by the class probabilities, can be reliable criteria for assessing whether the input label is the true label or the corrupted one. In this work, we exploit this observation and propose a novel discriminator metric called confidence error and a sieving strategy called CONFES to differentiate between the clean and noisy samples effectively. We provide theoretical guarantees on the probability of error for our proposed metric. Then, we experimentally illustrate the superior performance of our proposed approach compared to recent studies on various settings, such as synthetic and real-world label noise. Moreover, we show CONFES can be combined with other state-of-the-art approaches, such as Co-teaching and DivideMix to further improve model performance.

[1]  Florian Tramèr,et al.  Quantifying Memorization Across Neural Language Models , 2022, ICLR.

[2]  Tongliang Liu,et al.  Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotations , 2021, ICLR.

[3]  Mingming Gong,et al.  Instance-dependent Label-noise Learning under a Structural Causal Model , 2021, NeurIPS.

[4]  Daniel Coelho de Castro,et al.  Active label cleaning for improved dataset quality under resource constraints , 2021, Nature Communications.

[5]  Erkun Yang,et al.  Understanding and Improving Early Stopping for Learning with Noisy Labels , 2021, NeurIPS.

[6]  Hanlin Tang,et al.  On the geometry of generalization and memorization in deep neural networks , 2021, ICLR.

[7]  Qi Wu,et al.  Jo-SRC: A Contrastive Approach for Combating Noisy Labels , 2021, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[8]  Se-Young Yun,et al.  FINE Samples for Learning with Noisy Labels , 2021, NeurIPS.

[9]  Samy Bengio,et al.  Understanding deep learning (still) requires rethinking generalization , 2021, Commun. ACM.

[10]  Gang Niu,et al.  Learning Noise Transition Matrix from Only Noisy Labels via Total Variation Regularization , 2021, ICML.

[11]  Pheng-Ann Heng,et al.  Robustness of Accuracy Metric and its Inspirations in Learning with Noisy Labels , 2020, AAAI.

[12]  Yang Liu,et al.  Learning with Instance-Dependent Label Noise: A Sample Sieve Approach , 2020, ICLR.

[13]  Fan Zhang,et al.  P-DIFF: Learning Classifier with Noisy Labels based on Probability Difference Distributions , 2020, 2020 25th International Conference on Pattern Recognition (ICPR).

[14]  Hwanjun Song,et al.  Learning From Noisy Labels With Deep Neural Networks: A Survey , 2020, IEEE Transactions on Neural Networks and Learning Systems.

[15]  Dimitris N. Metaxas,et al.  Error-Bounded Correction of Noisy Labels , 2020, ICML.

[16]  Sheng Liu,et al.  Early-Learning Regularization Prevents Memorization of Noisy Labels , 2020, NeurIPS.

[17]  Gang Niu,et al.  Dual T: Reducing Estimation Error for Transition Matrix in Label-noise Learning , 2020, NeurIPS.

[18]  Gang Niu,et al.  Parts-dependent Label Noise: Towards Instance-dependent Label Noise , 2020, ArXiv.

[19]  Xiaohua Zhai,et al.  Are we done with ImageNet? , 2020, ArXiv.

[20]  Lei Feng,et al.  Combating Noisy Labels by Agreement: A Joint Training Method with Co-Regularization , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[21]  Junnan Li,et al.  DivideMix: Learning with Noisy Labels as Semi-supervised Learning , 2020, ICLR.

[22]  Kilian Q. Weinberger,et al.  Identifying Mislabeled Data using the Area Under the Margin Ranking , 2020, NeurIPS.

[23]  David F. Steiner,et al.  Chest Radiograph Interpretation with Deep Learning Models: Assessment with Radiologist-adjudicated Reference Standards and Population-adjusted Evaluation. , 2019, Radiology.

[24]  Weilong Yang,et al.  Beyond Synthetic Noise: Deep Learning on Controlled Noisy Labels , 2019, ICML.

[25]  Tiberiu T. Cocias,et al.  A survey of deep learning techniques for autonomous driving , 2019, J. Field Robotics.

[26]  E. Topol,et al.  A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis. , 2019, The Lancet. Digital health.

[27]  Gang Niu,et al.  Confidence Scores Make Instance-dependent Label-noise Learning Possible , 2019, ICML.

[28]  Jae-Gil Lee,et al.  Prestopping: How Does Early Stopping Help Generalization against Label Noise? , 2019, ArXiv.

[29]  Thomas L. Griffiths,et al.  Human Uncertainty Makes Classification More Robust , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[30]  Gang Niu,et al.  Are Anchor Points Really Indispensable in Label-Noise Learning? , 2019, NeurIPS.

[31]  Jae-Gil Lee,et al.  SELFIE: Refurbishing Unclean Samples for Robust Deep Learning , 2019, ICML.

[32]  Xingrui Yu,et al.  How does Disagreement Help Generalization against Label Corruption? , 2019, ICML.

[33]  Geraint Rees,et al.  Clinically applicable deep learning for diagnosis and referral in retinal disease , 2018, Nature Medicine.

[34]  Masashi Sugiyama,et al.  Co-teaching: Robust training of deep neural networks with extremely noisy labels , 2018, NeurIPS.

[35]  Gary Marcus,et al.  Deep Learning: A Critical Appraisal , 2018, ArXiv.

[36]  Li Fei-Fei,et al.  MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels , 2017, ICML.

[37]  Geoffrey E. Hinton,et al.  Dynamic Routing Between Capsules , 2017, NIPS.

[38]  Hongyi Zhang,et al.  mixup: Beyond Empirical Risk Minimization , 2017, ICLR.

[39]  Yoshua Bengio,et al.  A Closer Look at Memorization in Deep Networks , 2017, ICML.

[40]  Shai Shalev-Shwartz,et al.  Decoupling "when to update" from "how to update" , 2017, NIPS.

[41]  Richard Nock,et al.  Making Deep Neural Networks Robust to Label Noise: A Loss Correction Approach , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[42]  Frank Hutter,et al.  SGDR: Stochastic Gradient Descent with Warm Restarts , 2016, ICLR.

[43]  Jian Sun,et al.  Identity Mappings in Deep Residual Networks , 2016, ECCV.

[44]  Xiaogang Wang,et al.  Learning from massive noisy labeled data for image classification , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[45]  Nagarajan Natarajan,et al.  Learning with Noisy Labels , 2013, NIPS.

[46]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[47]  Fei-Fei Li,et al.  ImageNet: A large-scale hierarchical image database , 2009, 2009 IEEE Conference on Computer Vision and Pattern Recognition.

[48]  Chen Gong,et al.  Robust early-learning: Hindering the memorization of noisy labels , 2021, ICLR.

[49]  Pheng-Ann Heng,et al.  Noise against noise: stochastic label noise helps combat inherent label noise , 2021, ICLR.

[50]  Yang Liu,et al.  The importance of understanding instance-level noisy labels , 2021, ArXiv.

[51]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[52]  Alex Krizhevsky,et al.  Learning Multiple Layers of Features from Tiny Images , 2009 .