Metraipolis › archive › dc_db8deb24ded9

Season: Spectrum-Aware Orthogonal Gradient Refinement for Transfer-Based Adversarial Attacks — arXiv:2608.04441

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.04441
License (as archived)
CC0-1.0
Archived
2026-10-31T11:27:36.533Z
SHA-256 of the archived text
63047e31b40e1f8b0d402643b8e33970bb2acd8893656d76de0980e213aac47e
Stable URL of this page
https://metraipolis.com/documento/dc_db8deb24ded9

How to cite

Season: Spectrum-Aware Orthogonal Gradient Refinement for Transfer-Based Adversarial Attacks — arXiv:2608.04441 — arXiv (en), https://arxiv.org/abs/2608.04441 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_db8deb24ded9 (Metraipolis archive dc_db8deb24ded9, sha256 63047e31b40e1f8b0d402643b8e33970bb2acd8893656d76de0980e213aac47e)
BibTeX
@misc{dcdb8deb24ded9,
  title = {Season: Spectrum-Aware Orthogonal Gradient Refinement for Transfer-Based Adversarial Attacks — arXiv:2608.04441},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.04441},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_db8deb24ded9; sha256 63047e31b40e1f8b0d402643b8e33970bb2acd8893656d76de0980e213aac47e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.04441. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Transfer-based adversarial attacks often transfer poorly across heterogeneous architectures because CNNs favor local textures while Vision Transformers (ViTs) rely on global shapes. We propose Season, a spectrum-aware orthogonal gradient refinement framework for L-infinity transfer attacks against black-box target models on ImageNet, using a white-box surrogate. Season decomposes each update into a low-frequency branch capturing structural cues and a high-frequency branch capturing textures. A low-saliency guidance scheme reallocates high-frequency energy to background regions, preserving foreground structures that ViTs depend on. An orthogonal projection then forces the textural update to lie in the orthogonal complement of the structural direction, mitigating feature interference. As a training-free plug-and-play wrapper, Season enhances eight gradient-stabilization and input-enhancement attacks without modifying their cores. Across eight CNN, ViT, and MLP targets, Season improves transfer success rate by 6.6 percentage points on average and up to 16.0 points over strong baselines under a unified protocol.

Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-08-06 in cs.CR).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2608.04441 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-31T11:17:19.459Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-08-06?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).