Enhancing Adversarial Robustness via Score-Based Optimization

This repository contains the code for the paper "Enhancing Adversarial Robustness via Score-Based Optimization" (NeurIPS 2023) in Python.

Pre-train classifier models

Please refer to https://github.com/meliketoy/wide-resnet.pytorch

Pre-train diffusion models

Please refer to https://github.com/NVlabs/edm and https://github.com/openai/guided-diffusion.

Test-time defense and robustness evaluation

Please first put all the datasets in the datasets directory and all the pre-trained models in the pretrained_models directory.

To reproduce the experimental results in the main text, please run:

bash run.sh