An Improved Method of Identifying Mislabeled Data and the Mislabeled Data in MNIST and CIFAR-10 Appendix Findings in Fashion-MNIST
8 Pages Posted: 12 Jan 2018
Date Written: January 5, 2018
Abstract
Objects classification is an important part of machine learning and the quality of the training data plays an important role in it. Some mislabeled data detection techniques have been proposed; however, there is no such work done on MNIST and CIFAR-10, the result on which is an important criterion for a machine learning models or algorithms. In this paper I develop an improved method to identify mislabeled data and find 675 mislabeled data in MNIST, 118 mislabeled data in CIFAR-10, some mislabeled data in fashion MNIST.
Suggested Citation: Suggested Citation
Zhang, Xinbin, An Improved Method of Identifying Mislabeled Data and the Mislabeled Data in MNIST and CIFAR-10 Appendix Findings in Fashion-MNIST (January 5, 2018). Available at SSRN: https://ssrn.com/abstract=3097307 or http://dx.doi.org/10.2139/ssrn.3097307
Do you have negative results from your research you’d like to share?
Feedback
Feedback to SSRN
If you need immediate assistance, call 877-SSRNHelp (877 777 6435) in the United States, or +1 212 448 2500 outside of the United States, 8:30AM to 6:00PM U.S. Eastern, Monday - Friday.