Metraipolis › archive › dc_78f10d0442db
Decodable but Not Accessible: Auditing Distance-Based Reliability Estimation on Disentangled Skin-Lesion Representations — arXiv:2608.11267
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.11267
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T05:49:45.919Z
- SHA-256 of the archived text
7ed4b53997545e4a70a584298aa846cabbc3f7612ea9b4be6b8429a551c84357- Stable URL of this page
https://metraipolis.com/documento/dc_78f10d0442db
How to cite
Decodable but Not Accessible: Auditing Distance-Based Reliability Estimation on Disentangled Skin-Lesion Representations — arXiv:2608.11267 — arXiv (en), https://arxiv.org/abs/2608.11267 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_78f10d0442db (Metraipolis archive dc_78f10d0442db, sha256 7ed4b53997545e4a70a584298aa846cabbc3f7612ea9b4be6b8429a551c84357)
BibTeX
@misc{dc78f10d0442db,
title = {Decodable but Not Accessible: Auditing Distance-Based Reliability Estimation on Disentangled Skin-Lesion Representations — arXiv:2608.11267},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.11267},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_78f10d0442db; sha256 7ed4b53997545e4a70a584298aa846cabbc3f7612ea9b4be6b8429a551c84357},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.11267. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Distance-based reliability estimation assumes that a representation's geometry reflects its trustworthiness, yet this assumption is rarely tested under training interventions that reshape geometry directly. We audit this assumption under domain-adversarial representation learning using a disentanglement dose-response ladder. Three checkpoint families share the same architecture and a 16-dimensional representation, differing only in orthogonality strength (lambda = 0, 1, 5). Representation geometry changed substantially with disentanglement strength: the condition number shifted by two orders of magnitude (Kendall tau = 0.84, exact p = 2.8e-5). This change was not accompanied by improved reliability estimation: Mahalanobis-distance AUROC (ISIC-test vs. PAD-UFES) remained flat and below chance (about 0.40) at every level, with no significant association with any of five geometry metrics tested. The same failure was observed for cosine-to-centroid and pooled k-nearest-neighbor scorers, plus three non-distance-based scorers: an energy-based confidence score, Virtual-Logit Matching, and a kernel density estimator. Seven of eight scorers converged on the same result; the energy-based score showed an isolated upward trend that we report but do not treat as evidence against the overall pattern. A supervised probe with no access to the training objective recovered domain membership from the identical embeddings at 0.72-0.81 AUROC across every level, showing that the relevant information was not absent from the representation. These findings indicate that classification performance alone can overlook whether information in a learned representation is organized in a form that downstream reliability estimators can use. Information can remain decodable while becoming largely inaccessible to non-probing reliability estimators. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-08-13 in cs.CV). Categorie arXiv: Image and Video Processing (eess.IV) ; Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2608.11267 — 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 2026-10-19T05:42:25.931Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-08-13?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).