Search Results for author:
Found 12 papers, 8 papers with code
What do Vision Transformers Learn? A Visual Exploration
In addition, we show that ViTs maintain spatial information in all layers except the final layer.
Improving Robustness with Adaptive Weight Decay
We propose adaptive weight decay, which automatically tunes the hyper-parameter for weight decay during each training iteration.
Plug-In Inversion: Model-Agnostic Inversion for Vision with Data Augmentations
Existing techniques for model inversion typically rely on hard-to-tune regularizers, such as total variation or feature regularization, which must be individually calibrated for each network in order to produce adequate images.
DP-InstaHide: Data Augmentations Provably Enhance Guarantees Against Dataset Manipulations
Data poisoning and backdoor attacks manipulate training data to induce security breaches in a victim model.
DP-InstaHide: Provably Defusing Poisoning and Backdoor Attacks with Differentially Private Data Augmentations
The InstaHide method has recently been proposed as an alternative to DP training that leverages supposed privacy properties of the mixup augmentation, although without rigorous guarantees.
The Uncanny Similarity of Recurrence and Depth
It is widely believed that deep neural networks contain layer specialization, wherein neural networks extract hierarchical features representing edges and patterns in shallow layers and complete objects in deeper layers.
Strong Data Augmentation Sanitizes Poisoning and Backdoor Attacks Without an Accuracy Tradeoff
Data poisoning and backdoor attacks manipulate victim models by maliciously modifying training data.
Towards Accurate Quantization and Pruning via Data-free Knowledge Transfer
Further, we demonstrate that the compact structure and corresponding initialization from the Lottery Ticket Hypothesis can also help in data-free training.
Breaking certified defenses: Semantic adversarial examples with spoofed robustness certificates
To deflect adversarial attacks, a range of "certified" classifiers have been proposed.
Label Smoothing and Logit Squeezing: A Replacement for Adversarial Training?
Adversarial training is one of the strongest defenses against adversarial attacks, but it requires adversarial examples to be generated for every mini-batch during optimization.
Adversarially robust transfer learning
By training classifiers on top of these feature extractors, we produce new models that inherit the robustness of their parent networks.
Adversarial Training for Free!
Adversarial training, in which a network is trained on adversarial examples, is one of the few defenses against adversarial attacks that withstands strong attacks.
