An Effective Baseline for Robustness to Distributional Shift

Refraining from confidently predicting when faced with categories of inputs different from those seen during training is an important requirement for the safe deployment of deep learning systems. While simple to state, this has been a particularly challenging problem in deep learning, where models often end up making overconfident predictions in such situations. In this work we present a simple, but highly effective approach to deal with out-of-distribution detection that uses the principle of abstention: when encountering a sample from an unseen class, the desired behavior is to abstain from predicting. Our approach uses a network with an extra abstention class and is trained on a dataset that is augmented with an uncurated set consisting of a large number of out-of-distribution (OoD) samples that are assigned the label of the abstention class; the model is then trained to learn an effective discriminator between in- and out-of-distribution samples.

Paper (arXiv) · IEEE Xplore

BibTeX

@inproceedings{thulasidasan2021effective,
  title     = {An Effective Baseline for Robustness to Distributional Shift},
  author    = {Thulasidasan, Sunil and Thapa, Sushil and Dhaubhadel, Sayera and Chennupati, Gopinath and Bhattacharya, Tanmoy and Bilmes, Jeff},
  booktitle = {2021 20th IEEE International Conference on Machine Learning and Applications (ICMLA)},
  pages     = {278--285},
  year      = {2021},
  organization = {IEEE},
  doi       = {10.1109/ICMLA52953.2021.00050}
}