AI Paper Atlas

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

Overcoming the Convex Barrier for Simplex Inputs

Harkirat Singh, M. Pawan Kumar, Philip Torr, Krishnamurthy Dj Dvijotham

NeurIPS 2021 · Adversarial robustness and security › Certified and empirical robustness · approach: Neural network robustness verification · 0 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
use Simplex relaxation instead of Planet relaxation, both solved exactly with Gurobi (ablation)Gurobi · MNIST OVAL Wideverified accuracy: 31.7% -> 45.2%Improved
Gurobi Simplex vs Gurobi Planet (ablation)Gurobi · CIFAR-10 OVAL Wideverified accuracy: 18.4% -> 29.4%Improved
Simplex Verify instead of Opt-Lirpa Planet at same timeLiRPA solver · CIFAR-10 VNN Medverified accuracy: 48.8% -> 59.4%Improved
Simplex Verify instead of Opt-Lirpa PlanetLiRPA solver · CIFAR-10 VNN Bigverified accuracy: 60.0% -> 66.4%Improved
Simplex Verify instead of Opt-Lirpa PlanetLiRPA solver · MNIST OVAL Wideverified accuracy: 31.0% -> 44.6%Improved
Simplex Verify instead of Opt-Lirpa Planet on SGD-trained modelLiRPA solver · CIFAR-10 OVAL Wide (SGD)verified accuracy: 13.5% -> 22.4%Improved
Simplex Verify instead of Opt-Lirpa Planet on small-margin Deep modelLiRPA solver · CIFAR-10 OVAL Deepverified accuracy: 10.8% -> 13.4%Improved
fast LiRPA solver for Simplex relaxation instead of GurobiSimplex Verify · MNIST OVAL Wideverified accuracy: 45.2% -> 44.6%No clear effect
fast LiRPA solver for Simplex relaxation instead of GurobiSimplex Verify · MNIST OVAL Widetime per sample: 72.47s -> 0.04sImproved
Simplex Verify vs Opt-Lirpa Planet runtimeLiRPA solvers · CIFAR-10 OVAL Deeptime per sample: 0.06s -> 0.05sNo clear effect
Simplex Verify instead of Opt-Lirpa Planet, text-perturbation, network 1ConcatBOW · Food-101 10-classverified accuracy: 17.9% -> 42.2%Improved
Simplex Verify instead of Opt-Lirpa Planet, text-perturbation, network 2ConcatBOW · Food-101 10-classverified accuracy: 08.0% -> 32.8%Improved

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