Metraipolis › archive › dc_b04004b91dcc
Augmenting Visual Anomaly Detection with Automated Interpretability — arXiv:2609.33818
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.33818
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T21:01:48.994Z
- SHA-256 of the archived text
44ab58243f5df3fb6bd2b72becea7b5bc4923ca02af25374823f959168d87599- Stable URL of this page
https://metraipolis.com/documento/dc_b04004b91dcc
How to cite
Augmenting Visual Anomaly Detection with Automated Interpretability — arXiv:2609.33818 — arXiv (en), https://arxiv.org/abs/2609.33818 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b04004b91dcc (Metraipolis archive dc_b04004b91dcc, sha256 44ab58243f5df3fb6bd2b72becea7b5bc4923ca02af25374823f959168d87599)
BibTeX
@misc{dcb04004b91dcc,
title = {Augmenting Visual Anomaly Detection with Automated Interpretability — arXiv:2609.33818},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.33818},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b04004b91dcc; sha256 44ab58243f5df3fb6bd2b72becea7b5bc4923ca02af25374823f959168d87599},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33818. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Visual anomaly detectors identify deviations from known-normal data, but their anomaly signals may mix evidence of actual anomalies with benign visual variation. We investigate whether automated interpretability can augment visual anomaly detectors by identifying and intervening on different components of this signal. We decompose PatchCore nearest-normal residuals into sparse features using Sparse Autoencoders (SAEs), and provide high-activation and contrastive non-active examples to a Multimodal LLM, which describes each feature and labels it as anomaly, distractor, or uncertain. These labels guide interventions in the SAE hidden representation, where distractor features are suppressed and anomaly features amplified. The edited representation is then used to reconstruct patch embeddings, which are rescored with PatchCore. Across 40 categories from four benchmarks, applying both interventions jointly improves macro-average image-level AUROC from 0.8724 to 0.8857 on source data and from 0.8066 to 0.8210 under synthetic corruptions. On three additional RobustAD categories with real acquisition shifts, the same interventions improve AUROC from 0.8745 to 0.9056 on source data and from 0.6069 to 0.6599 under real acquisition shifts. Finally, individual feature interventions across all 43 categories show that the MLLM labels are aligned in aggregate with how features differently affect normal and anomalous images. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.33818 — 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 2026-10-07T20:29:02.759Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).