All AI Labs Business News Newsletters Research Safety Tools Topics Sources
🔬 Research 9 min read · 6 years ago

A Discussion of 'Adversarial Examples Are Not Bugs, They Are Features': Adversarial Example Researchers Need to Expand What is Meant by 'Robustness'

#adversarial examples #robustness #distribution shift #machine learning #neural networks
A Discussion of 'Adversarial Examples Are Not Bugs, They Are Features': Adversarial Example Researchers Need to Expand What is Meant by 'Robustness'
Curated from Distill.pub Read original →

DeepTrendLab's Take on A Discussion of 'Adversarial Examples Are Not Bugs, They...

Adversarial example researchers should expand their focus beyond gradient perturbations to connect with distributional robustness literature, recognizing that models' vulnerability stems from relying on superficial statistical correlations.

This article was originally published on Distill.pub. Read the full piece at the source.

Read full article on Distill.pub →

DeepTrendLab curates AI news from 50+ sources. All original content and rights belong to Distill.pub. DeepTrendLab's analysis is independently written and does not represent the views of the original publisher.

More Research