Metraipolis › archive › dc_374f10f8f352
Beyond Scarcity: When Encoder Adaptation Is Worthwhile — arXiv:2610.41213
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.41213
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T08:11:46.874Z
- SHA-256 of the archived text
0625d84f11811cb0f341788f9c7deb223e054328cd7046020f18b9cb04d804b5- Stable URL of this page
https://metraipolis.com/documento/dc_374f10f8f352
How to cite
Beyond Scarcity: When Encoder Adaptation Is Worthwhile — arXiv:2610.41213 — arXiv (en), https://arxiv.org/abs/2610.41213 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_374f10f8f352 (Metraipolis archive dc_374f10f8f352, sha256 0625d84f11811cb0f341788f9c7deb223e054328cd7046020f18b9cb04d804b5)
BibTeX
@misc{dc374f10f8f352,
title = {Beyond Scarcity: When Encoder Adaptation Is Worthwhile — arXiv:2610.41213},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.41213},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_374f10f8f352; sha256 0625d84f11811cb0f341788f9c7deb223e054328cd7046020f18b9cb04d804b5},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41213. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Long-tailed recognition often learns a shared representation before correcting the classifier for class imbalance. Once the classifier has been rebalanced, however, final accuracy does not reveal whether further encoder adaptation remains worthwhile. Standard imbalance benchmarks obscure this question because they jointly vary the number of distinct examples available to rare classes and the frequency with which those examples shape the representation. We introduce a controlled evaluation framework that separates unique support from representation allocation and estimates the incremental value of encoder updates through matched classifier-only and full-network continuations from a common corrected model. The framework combines stage-specific controls, multiple representation and adaptation objectives, whole-distribution metrics, and hierarchical inference over class assignments, support reconstructions, and optimization runs. Across the controlled study, the same low-support regime moves from one in which classifier-only training is adequate to one in which encoder adaptation yields a practically meaningful benefit solely as representation allocation becomes more imbalanced. The transition persists across classifier corrections, representation objectives, normalization and augmentation controls, and ImageNet-scale evaluation, while preserving frequent-class accuracy. Under the prespecified one-percentage-point operating margin, the two regimes are separated by simultaneous confidence bounds. These results show that benchmark construction can alter a post-correction training decision and that support size alone does not characterize the remaining value of representation adaptation. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2610.41213 — 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 2026-11-03T08:02:38.911Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).