Research map / Adversarial robustness and security / Black-box and transfer attacks

Training Robust Ensembles Requires Rethinking Lipschitz Continuity

ICLR 2025 · Adversarial robustness and security › Black-box and transfer attacks · approach: Transferable adversarial examples

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Studies adversarial robustness of model ensembles against transferable adversarial examples, arguing that ensemble diversity alone is insufficient and that Lipschitz continuity should be rethought when training robust ensembles.

Problem

Adversarial transferability endangers classifiers, including black-box ones; ensembles resist transfer only if adversarial examples rarely fool most members, and prior work focuses mainly on improving ensemble diversity.

Approach

Based only on the truncated abstract: argues that the focus on diversity is insufficient and that the Lipschitz continuity of ensemble members should be rethought as the key property for training robust ensembles.

Closest papers by meaning

• Certifying Ensembles: A General Certification Theory with S-Lipschitzness (ICML 2023 · similarity 0.90)
• Improving Ensemble Robustness by Collaboratively Promoting and Demoting Adversarial Robustness (AAAI 2021 · similarity 0.90)
• Rethinking Model Ensemble in Transfer-based Adversarial Attacks (ICLR 2024 · similarity 0.89)
• Understanding Model Ensemble in Transferable Adversarial Attack (ICML 2025 · similarity 0.88)
• A Recipe for Improved Certifiable Robustness (ICLR 2024 · similarity 0.88)
• Some Fundamental Aspects about Lipschitz Continuity of Neural Networks (ICLR 2024 · similarity 0.88)