From Noisy Prediction to True Label: Noisy Prediction Calibration via Generative Model

Heesun Bae, Seungjae Shin, Byeonghu Na, Joonho Jang, Kyungwoo Song, Il-Chul Moon
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:1277-1297, 2022.

Abstract

Noisy labels are inevitable yet problematic in machine learning society. It ruins the generalization of a classifier by making the classifier over-fitted to noisy labels. Existing methods on noisy label have focused on modifying the classifier during the training procedure. It has two potential problems. First, these methods are not applicable to a pre-trained classifier without further access to training. Second, it is not easy to train a classifier and regularize all negative effects from noisy labels, simultaneously. We suggest a new branch of method, Noisy Prediction Calibration (NPC) in learning with noisy labels. Through the introduction and estimation of a new type of transition matrix via generative model, NPC corrects the noisy prediction from the pre-trained classifier to the true label as a post-processing scheme. We prove that NPC theoretically aligns with the transition matrix based methods. Yet, NPC empirically provides more accurate pathway to estimate true label, even without involvement in classifier learning. Also, NPC is applicable to any classifier trained with noisy label methods, if training instances and its predictions are available. Our method, NPC, boosts the classification performances of all baseline models on both synthetic and real-world datasets. The implemented code is available at https://github.com/BaeHeeSun/NPC.

Cite this Paper

BibTeX
@InProceedings{pmlr-v162-bae22a, title = {From Noisy Prediction to True Label: Noisy Prediction Calibration via Generative Model}, author = {Bae, Heesun and Shin, Seungjae and Na, Byeonghu and Jang, Joonho and Song, Kyungwoo and Moon, Il-Chul}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {1277--1297}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/bae22a/bae22a.pdf}, url = {https://proceedings.mlr.press/v162/bae22a.html}, abstract = {Noisy labels are inevitable yet problematic in machine learning society. It ruins the generalization of a classifier by making the classifier over-fitted to noisy labels. Existing methods on noisy label have focused on modifying the classifier during the training procedure. It has two potential problems. First, these methods are not applicable to a pre-trained classifier without further access to training. Second, it is not easy to train a classifier and regularize all negative effects from noisy labels, simultaneously. We suggest a new branch of method, Noisy Prediction Calibration (NPC) in learning with noisy labels. Through the introduction and estimation of a new type of transition matrix via generative model, NPC corrects the noisy prediction from the pre-trained classifier to the true label as a post-processing scheme. We prove that NPC theoretically aligns with the transition matrix based methods. Yet, NPC empirically provides more accurate pathway to estimate true label, even without involvement in classifier learning. Also, NPC is applicable to any classifier trained with noisy label methods, if training instances and its predictions are available. Our method, NPC, boosts the classification performances of all baseline models on both synthetic and real-world datasets. The implemented code is available at https://github.com/BaeHeeSun/NPC.} }
Endnote
%0 Conference Paper %T From Noisy Prediction to True Label: Noisy Prediction Calibration via Generative Model %A Heesun Bae %A Seungjae Shin %A Byeonghu Na %A Joonho Jang %A Kyungwoo Song %A Il-Chul Moon %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-bae22a %I PMLR %P 1277--1297 %U https://proceedings.mlr.press/v162/bae22a.html %V 162 %X Noisy labels are inevitable yet problematic in machine learning society. It ruins the generalization of a classifier by making the classifier over-fitted to noisy labels. Existing methods on noisy label have focused on modifying the classifier during the training procedure. It has two potential problems. First, these methods are not applicable to a pre-trained classifier without further access to training. Second, it is not easy to train a classifier and regularize all negative effects from noisy labels, simultaneously. We suggest a new branch of method, Noisy Prediction Calibration (NPC) in learning with noisy labels. Through the introduction and estimation of a new type of transition matrix via generative model, NPC corrects the noisy prediction from the pre-trained classifier to the true label as a post-processing scheme. We prove that NPC theoretically aligns with the transition matrix based methods. Yet, NPC empirically provides more accurate pathway to estimate true label, even without involvement in classifier learning. Also, NPC is applicable to any classifier trained with noisy label methods, if training instances and its predictions are available. Our method, NPC, boosts the classification performances of all baseline models on both synthetic and real-world datasets. The implemented code is available at https://github.com/BaeHeeSun/NPC.
APA
Bae, H., Shin, S., Na, B., Jang, J., Song, K. & Moon, I.. (2022). From Noisy Prediction to True Label: Noisy Prediction Calibration via Generative Model. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:1277-1297 Available from https://proceedings.mlr.press/v162/bae22a.html.

Related Material