SelectiveNet: A Deep Neural Network with an Integrated Reject Option
This addresses the need for reliable uncertainty estimation in deep learning applications, offering a novel approach to selective prediction that outperforms existing methods.
The paper tackled the problem of selective prediction in deep neural networks by introducing SelectiveNet, an architecture with an integrated reject option trained end-to-end for simultaneous optimization of classification/regression and rejection, resulting in consistently improved risk-coverage trade-offs and new state-of-the-art results on several datasets.
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. Existing rejection mechanisms are based mostly on a threshold over the prediction confidence of a pre-trained network. In contrast, SelectiveNet is trained to optimize both classification (or regression) and rejection simultaneously, end-to-end. The result is a deep neural network that is optimized over the covered domain. In our experiments, we show a consistently improved risk-coverage trade-off over several well-known classification and regression datasets, thus reaching new state-of-the-art results for deep selective classification.