AI Paper Atlas

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

Robust-by-Design Classification via Unitary-Gradient Neural Networks

Fabio Brau, Giulio Rossolini, Alessandro Biondi, Giorgio Buttazzo

AAAI 2023 · Adversarial robustness and security › Certified and empirical robustness · approach: Neural network robustness verification · 4 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 unbounded rather than bounded UPD (ablation)UGNN · CIFAR10 testaccuracy: UGNN+Abs: 71.9±0.29 -> 72.1±0.54 (Std.Norm); 69.2±0.31 -> 68.9±0.81 (Raw). UGNN+MaxMin: 72.6±0.79 -> 72.7±0.38 (Std.NormMixed
Apply standard normalization to inputs (ablation)UGNN · CIFAR10 testaccuracy: UGNN+Abs+updB: 69.2±0.31 -> 71.9±0.29; UGNN+MaxMin+updB: 70.4±0.52 -> 72.6±0.79; UGNN+OPLU+updB: 70.5±0.39 -> 71.9±0.09Mixed
Replace MaxMin or OPLU with abs activation (ablation)UGNN · CIFAR10 testaccuracy: Std.Norm: 71.9±0.09 to 72.1±0.54 (OPLU to Abs, updU); Raw: 70.6±0.45 to 68.9±0.81 (OPLU to Abs, updU)No clear effect
Increase input size from 32 to 256 (ablation)UGNN+MaxMin+updB · CIFAR10, Rawaccuracy: 70.4±0.52 -> 78.4±0.29Improved
Increase input size from 32 to 256 (ablation)UGNN+MaxMin+updB · CIFAR10, Std.Normaccuracy: 72.6±0.79 -> 76.5±0.35Improved
Use the UGNN lower bound instead of the compared 1-Lipschitz models’ lower boundUGNN · CIFAR10, RawLB/MAP: ResNet9 .34±.063; LargeConvNet .46±.057; LipConvNet5 .58±.069; UGNN+MaxMin+updU .71±.088Improved
Use the UGNN lower bound instead of the compared 1-Lipschitz models’ lower boundUGNN · CIFAR10, Std.NormLB/MAP: ResNet9 .26±.036; LargeConvNet .44±.027; LipConvNet5 .52±.031; UGNN+MaxMin+updB .96±.044Improved
Use UGNN+MaxMin+updB-256-raw rather than the 1-Lipschitz modelsUGNN · CIFAR10 certifiable ε-robust classificationsaccuracyImproved

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