Robustness of classifiers: from adversarial to random noise

Authors: Alhussein Fawzi, Seyed-Mohsen Moosavi-Dezfooli, Pascal Frossard

NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We perform experiments and show that the derived bounds provide very accurate estimates when applied to various state-of-the-art deep neural networks and datasets.
Researcher Affiliation Academia École Polytechnique Fédérale de Lausanne Lausanne, Switzerland {alhussein.fawzi, seyed.moosavi, pascal.frossard} at epfl.ch
Pseudocode No The paper does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statements about releasing source code for the methodology, nor does it include a link to a code repository.
Open Datasets Yes Le Net (MNIST), Le Net (CIFAR-10), VGG-F (Image Net), VGG-19 (Image Net)
Dataset Splits No The paper mentions 'test set' (D) for evaluating β(f; m), but it does not provide specific details on training/validation splits, percentages, or sample counts.
Hardware Specification Yes We also gratefully acknowledge the support of NVIDIA Corporation with the donation of the Tesla K40 GPU used for this research.
Software Dependencies No The paper does not specify any software dependencies with version numbers.
Experiment Setup No The paper evaluates pre-existing state-of-the-art classifiers (LeNet, VGG-F, VGG-19) on various datasets but does not provide specific details on hyperparameters, training configurations, or model initialization used for these classifiers in their experiments.