MadryLab
Towards a Principled Science of Deep Learning
Built from MadryLab's public GitHub profile, MIT. Not on Take My Repo yet, so there are no posts or updates here.
A library for experimenting with, training and evaluating neural networks, with a focus on adversarial robustness.
A challenge to explore adversarial robustness of neural networks on MNIST.
A challenge to explore adversarial robustness of neural networks on CIFAR10.
Attribute (or cite) statements generated by LLMs back to in-context information.
Datasets for the paper "Adversarial Examples are not Bugs, They Are Features"
Code for "Learning Perceptually-Aligned Representations via Adversarial Robustness"
Decomposing and Editing Predictions by Modeling Model Computation
Notebooks for reproducing the paper "Computer Vision with a Single (Robust) Classifier"
Code for "Prior Convictions: Black-Box Adversarial Attacks with Bandits and Priors"
Distilling Model Failures as Directions in Latent Space
Investigating the robustness of state-of-the-art CNN architectures to simple spatial transformations.
Dataset Interfaces: Diagnosing Model Failures Using Controllable Counterfactual Generation