Nonparametric Classification under Adversarial Attacks: Decision-Boundary Geometry and Optimal Rates
Source: arxiv.org/abs/2611.19347
Unlike in continuous regression, a small input perturbation in classification can move an input across a decision boundary and flip the predicted label. Decision-boundary geometry is therefore essential to understanding adversarial classification risk. Under anisotropic smoothness and margin conditions, we quantify this geometric contribution and establish sharp minimax rates for adversarial excess risk relative to the ordinary Bayes risk. These rates comprise a statistical term governed by the harmonic mean smoothness and a geometric cost of adversarial perturbations governed by the least…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2611.19347
curl -X POST /api/v1/records/arxiv/2611.19347/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2611.19347/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'