AI Paper Atlas

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

On the Sharp Input-Output Analysis of Nonlinear Systems under Adversarial Attacks

Jihun Kim, Yuchen Fang, Javad Lavaei

ICML 2026 · Adversarial robustness and security › Certified and empirical robustness · approach: Provable robust learning guarantees

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 l2-norm estimator instead of least squares under probabilistic adversarial attacksinput-output mapping estimator · synthetic tanh system (100 states, 5 inputs, 10 outputs)estimation error of G*Improved
Use l2-norm estimator instead of least squares on power-grid swing dynamics under attackinput-output mapping estimator · nonlinear swing dynamics power systemestimation errorImproved
Use non-Gaussian nonzero-mean inputs Unif[-8,10]^5 instead of Gaussian N(0,100 I_5) (ablation)l2-norm estimator · synthetic tanh systemestimation errorNo clear effect
Use l2-norm instead of l1 or l-infinity norm estimators (ablation)input-output mapping estimator · synthetic tanh systemestimation errorImproved
Increase input memory length tau (ablation)l2-norm estimator · synthetic tanh system, non-Gaussian inputseventual estimation errorImproved
Increase contraction factor rho (ablation)l2-norm estimator · synthetic tanh system, non-Gaussian inputsestimation errorHurt
Increase sample horizon T (ablation)l2-norm estimator · synthetic tanh systemestimation errorNo clear effect

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