To refer to this page use:
|Abstract:||The classification performance of deep neural networks has begun to asymptote at near-perfect levels. However, their ability to generalize outside the training set and their robustness to adversarial attacks have not. In this paper , we make progress on this problem by training with full label distributions that reflect human perceptual uncertainty. We first present a new benchmark dataset which we call CIFAR10H, containing a full distribution of human labels for each image of the CIFAR10 test set. We then show that, while contemporary classifiers fail to exhibit human-like uncertainty on their own, explicit training on our dataset closes this gap, supports improved generalization to increasingly out-of-training-distribution test datasets, and confers robustness to adversarial attacks.|
|Citation:||Peterson, Joshua, Battleday, Ruairidh, Griffiths, Thomas, and Russakovsky, Olga. (2019). Human uncertainty makes classification more robust. In Proceedings of the IEEE International Conference on Computer Vision, pp. 9617-9626.|
|Pages:||9617 - 9626|
|Type of Material:||Conference Article|
|Journal/Proceeding Title:||Proceedings of the IEEE International Conference on Computer Vision|
Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.