Luca Zappella
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Luca Zappella has a total of 4 co-authors. Here are the top co-authors by number of shared publications:
- Arno Blaas
University of Oxford
1 shared publication - Jason Ramapuram
University of Oxford
1 shared publication - Nicholas Apostoloff
Apple
1 shared publication - Xavier Suau
University of Oxford
1 shared publication
Latest Publications
- Open Access
Challenges of Adversarial Image Augmentations
Image augmentations applied during training are crucial for the generalization performance of image classifiers. Therefore, a large body of research has focused on finding the optimal augmentation policy for a given task. Yet, RandAugment [2], a simple random augmentation policy, has recently been shown to outperform existing sophisticated policies. Only Adversarial AutoAugment (AdvAA) [11], an approach based on the idea of adversarial training, has shown to be better than RandAugment. In this paper, we show that random augmentations are still competitive compared to an optimal adversarial approach, as well as to simple curricula, and conjecture that the success of AdvAA is due to the stochasticity of the policy controller network, which introduces a mild form of curriculum.
Open AccessChallenges of Adversarial Image Augmentations
Image augmentations applied during training are crucial for the generalization performance of image classifiers. Therefore, a large body of research has focused on finding the optimal augmentation policy for a given task. Yet, RandAugment [2], a simple random augmentation policy, has recently been shown to outperform existing sophisticated policies. Only Adversarial AutoAugment (AdvAA) [11], an approach based on the idea of adversarial training, has shown to be better than RandAugment. In this paper, we show that random augmentations are still competitive compared to an optimal adversarial approach, as well as to simple curricula, and conjecture that the success of AdvAA is due to the stochasticity of the policy controller network, which introduces a mild form of curriculum.