Code for our ICLR Trustworthy ML 2020 workshop paper "Improved Image Wasserstein Attacks and Defenses" - edwardjhu/improved_wasserstein

Verification of neural networks enables us to gauge their robustness against adversarial attacks. Verification algorithms fall into two categories: exact verifiers that run in exponential time and...

We introduce 3DB, an easy-to-use and extensible framework for debugging vision models with 3D rendering.

With little training data or compute, transfer learning is a simple way to obtain performant machine learning models. Learn how researchers at Microsoft research and MIT found adversarially robust ML...

Code for our NeurIPS 2019 *spotlight* "Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers" - GitHub - Hadisalman/smoothing-adversarial: Code for our Neu...

We present a randomized primal-dual algorithm that solves the problem minx maxy y^T A x to additive error epsilon in time nnz(A) + sqrt{nnz(A) n} / epsilon, for matrix A with larger dimension n and...

Raising the Cost of Malicious AI-Powered Image Editing - MadryLab/photoguard

Certified Patch Robustness via Smoothed Vision Transformers - MadryLab/smoothed-vit

Date: Oct 18th, 2021 Location: The workshop will be held virtually Zoom link: https://cmu.zoom.us/j/97234065013?pwd=VExidmdpaHJWODFLVEFjVGZEeUlXQT09 Registration link: https://formal-analysis.com/a...

Benchmark for LP-relaxed robustness verification of ReLU-networks - Hadisalman/robust-verify-benchmark

Transfer learning is a widely-used paradigm in deep learning, where models pre-trained on standard datasets can be efficiently adapted to downstream tasks. Typically, better pre-trained models...

Inspired by an episode of the Daily Show, we hacked together a technique for

Provable adversarial robustness at ImageNet scale. Contribute to locuslab/smoothing development by creating an account on GitHub.

Professional Conference Recording

Raising the Cost of Malicious AI-Powered Image Editing - MadryLab/photoguard

Over the last few years, the phenomenon of adversarial examples --- maliciously constructed inputs that fool trained machine learning models --- has captured the attention of the research...