AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

DVERGE: Diversifying Vulnerabilities for Enhanced Robust Generation of Ensembles

Huanrui Yang, Jingyang Zhang, Hongliang Dong, Nathan Inkawhich, Andrew Gardner, Andrew Touchet, Wesley Wilkes, Heath Berry, Hai Li

NeurIPS 2020 · Adversarial robustness and security › Certified and empirical robustness · approach: Attack evaluation and defense critique · 60 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train ensemble with DVERGE vulnerability diversification instead of standard trainingResNet-20 ensemble (3) · CIFAR-10 black-box eps=0.03accuracy (%): 0 -> 51.0Improved
use DVERGE instead of GALResNet-20 ensemble (3) · CIFAR-10 black-box eps=0.03accuracy (%): 39.8 -> 51.0Improved
use DVERGE vs baseline on clean dataResNet-20 ensemble (3) · CIFAR-10 cleanaccuracy (%): 93.9 -> 91.3Hurt
add more sub-models to DVERGE ensemble (3 to 8) (ablation)DVERGE · CIFAR-10 black-box eps=0.03accuracy (%): 51.0 -> 65.0Improved
add more sub-models to GAL ensemble (3 to 8)GAL · CIFAR-10 black-box eps=0.03accuracy (%): 39.8 -> 21.7Hurt
increase training eps from 0.03 to 0.07 (ablation)DVERGE (3 sub-models) · CIFAR-10 black-box eps=0.03accuracy (%): 4.2 -> 53.2Improved
increase training eps from 0.03 to 0.07 (ablation)DVERGE (3 sub-models) · CIFAR-10 cleanaccuracy (%): 92.9 -> 91.4Hurt
increase training eps from 0.05 to 0.07 (ablation)DVERGE (5 sub-models) · CIFAR-10 white-box eps=0.01accuracy (%): 48.9 -> 42.3Hurt
add adversarial training objective to DVERGE, compared with AdvT alone (ablation)DVERGE+AdvT (5) · CIFAR-10 cleanaccuracy (%): 78.6 -> 83.4Improved
add adversarial training objective to DVERGE, compared with AdvT alone (ablation)DVERGE+AdvT (5) · CIFAR-10 black-box eps=0.03accuracy (%): 73.8 -> 76.6Improved
add adversarial training objective to DVERGE, compared with AdvT alone (ablation)DVERGE+AdvT (5) · CIFAR-10 black-box eps=0.07accuracy (%): 63.3 -> 61.7Mixed
use DVERGE instead of adversarial training in computeDVERGE · CIFAR-10 training timehours: 11.5 -> 10.5Improved

Later papers that compared with this method

1 later accepted paper reports DVERGE/3 as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy