NIPS Proceedingsβ

Dan Hendrycks

1 Paper

  • Using Trusted Data to Train Deep Networks on Labels Corrupted by Severe Noise (2018)