Our method allows a user to set a desired risk level We consider the problem of selective prediction (also known as reject option) in deep neural networks, and introduce selectivenet, a deep neural architecture with an integrated reject option. At test time, the classifier rejects instances as needed, to grant the desired risk (with high probability).
I Nudes – Celeb Nudes
At test time, the classifier.
A deep neural network with an integrated reject option international conference on machine learning (icml 2019)
