Papers/2608.14594
🧪 Test?View on arXiv

Geometry Is Not Robustness: A Trajectory-Level Study of PGD Evaluation

Author1, Author2, Author3, Author4, Author5

adversarial robustnessPGDtrajectory analysismodel evaluation
2608.14594
Builder Relevance
70%
1h ago

Abstract

This paper investigates the effectiveness of trajectory-level diagnostics in evaluating adversarial robustness of models under PGD attacks.

Reality Card

Core Claim

Trajectory-level diagnostics do not independently measure adversarial robustness and should be used as complementary tools rather than replacements for standard robustness measurements.

Method / Result

Recorded full PGD trajectories across 3000 clean-correct samples per model.

Limitations

Trajectory metrics do not contribute equally to identifying robustness strength across different models.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.
← Back to all papers