# On the Adversarial Robustness of Neural Networks without Weight Transport

arXiv: https://arxiv.org/abs/1908.03560 · HTML: https://mlsanity.com/paper/1908.03560/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 Aug 2019.

## Built on by (most cited first, 5 of 5)

- [Benchmarking the Accuracy and Robustness of Feedback Alignment Algorithms](https://mlsanity.com/paper/2108.13446/) (2021)
- [Adversarial Robustness by Design through Analog Computing and Synthetic Gradients](https://mlsanity.com/paper/2101.02115/) (2021)
- [ROPUST: Improving Robustness through Fine-tuning with Photonic Processors and Synthetic Gradients](https://mlsanity.com/paper/2108.04217/) (2021)
- [Brain-like Flexible Visual Inference by Harnessing Feedback-Feedforward Alignment](https://mlsanity.com/paper/2310.20599/) (2023)
- [Intrinsic Biologically Plausible Adversarial Robustness](https://mlsanity.com/paper/2309.17348/) (2023)

## Abstract (arXiv)

Neural networks trained with backpropagation, the standard algorithm of deep learning which uses weight transport, are easily fooled by existing gradient-based adversarial attacks. This class of attacks are based on certain small perturbations of the inputs to make networks misclassify them. We show that less biologically implausible deep neural networks trained with feedback alignment, which do not use weight transport, can be harder to fool, providing actual robustness. Tested on MNIST, deep neural networks trained without weight transport (1) have an adversarial accuracy of 98% compared to 0.03% for neural networks trained with backpropagation and (2) generate non-transferable adversarial examples. However, this gap decreases on CIFAR-10 but is still significant particularly for small perturbation magnitude less than 1/2.