emaballarin / CARSO

👀🛡️ Code for the paper “Carefully Blending Adversarial Training and Purification Improves Adversarial Robustness” by Emanuele Ballarin, Alessio Ansuini and Luca Bortolussi (2024)

https://arxiv.org/abs/2306.06081

Geek Repo

Github PK Tool

adversarial-defense adversarial-machine-learning adversarial-robustness artificial-intelligence artificial-neural-networks cognitive-computing deep-learning deep-neural-networks machine-learning neural-network neural-networks neural-representation neural-representations neuro-inspired robust-machine-learning robustness variational-autoencoder variational-method

`CARSO`: Counter-Adversarial Recall of Synthetic Observations

Code for the paper “CARSO: Counter-Adversarial Recall of Synthetic Observations” by Emanuele Ballarin, Alessio Ansuini and Luca Bortolussi (2023).

Information about reproducibility of the results can be found in REPRO.md

Pre-trained models can be obtained by:

bash ./models/get_models_test.sh

About

👀🛡️ Code for the paper “Carefully Blending Adversarial Training and Purification Improves Adversarial Robustness” by Emanuele Ballarin, Alessio Ansuini and Luca Bortolussi (2024)

https://arxiv.org/abs/2306.06081

MIT License

Languages

Language:Python 82.9%Language:Shell 16.8%Language:Makefile 0.3%

CARSO: Counter-Adversarial Recall of Synthetic Observations

About

Languages

`CARSO`: Counter-Adversarial Recall of Synthetic Observations