Ian Goodfellow

TL;DR: It is found that adversarial examples span a contiguous subspace of large (~25) dimensionality, which indicates that it may be possible to design defenses against transfer-based attacks, even for models that are vulnerable to direct attacks.

TL;DR: This work introduces the first practical demonstration that cross-model transfer phenomenon enables attackers to control a remotely hosted DNN with no access to the model, its parameters, or its training data, and introduces the attack strategy of fitting a substitute model to the input-output pairs in this manner, then crafting adversarial examples based on this auxiliary model.

TL;DR: A number of empirical tests are proposed that directly measure the degree to which these learned features are invariant to different input transformations and find that stacked autoencoders learn modestly increasingly invariant features with depth when trained on natural images and convolutional deep belief networks learn substantially more invariant Features in each layer.

TL;DR: The Net2Net technique accelerates the experimentation process by instantaneously transferring the knowledge from a previous network to each new deeper or wider network, and demonstrates a new state of the art accuracy rating on the ImageNet dataset.

TL;DR: The core functionalities of the CleverHans library are presented, namely the attacks based on adversarial examples and defenses to improve the robustness of machine learning models to these attacks.