English

Training on test data: Removing near duplicates in Fashion-MNIST

Machine Learning 2019-06-21 v1 Computer Vision and Pattern Recognition Machine Learning

Abstract

MNIST and Fashion MNIST are extremely popular for testing in the machine learning space. Fashion MNIST improves on MNIST by introducing a harder problem, increasing the diversity of testing sets, and more accurately representing a modern computer vision task. In order to increase the data quality of FashionMNIST, this paper investigates near duplicate images between training and testing sets. Near-duplicates between testing and training sets artificially increase the testing accuracy of machine learning models. This paper identifies near-duplicate images in Fashion MNIST and proposes a dataset with near-duplicates removed.

Keywords

Cite

@article{arxiv.1906.08255,
  title  = {Training on test data: Removing near duplicates in Fashion-MNIST},
  author = {Christopher Geier},
  journal= {arXiv preprint arXiv:1906.08255},
  year   = {2019}
}