Metraipolis › archive › dc_076c2db315ab
Not Every Correction Helps: Gain-Guided Continual Test-Time Adaptation — arXiv:2609.36655
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.36655
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T18:38:21.177Z
- SHA-256 of the archived text
a9686deba193672ab58b45a6e8b0177a702eb4b77dcf3d95f50b11dc849331d1- Stable URL of this page
https://metraipolis.com/documento/dc_076c2db315ab
How to cite
Not Every Correction Helps: Gain-Guided Continual Test-Time Adaptation — arXiv:2609.36655 — arXiv (en), https://arxiv.org/abs/2609.36655 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_076c2db315ab (Metraipolis archive dc_076c2db315ab, sha256 a9686deba193672ab58b45a6e8b0177a702eb4b77dcf3d95f50b11dc849331d1)
BibTeX
@misc{dc076c2db315ab,
title = {Not Every Correction Helps: Gain-Guided Continual Test-Time Adaptation — arXiv:2609.36655},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.36655},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_076c2db315ab; sha256 a9686deba193672ab58b45a6e8b0177a702eb4b77dcf3d95f50b11dc849331d1},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36655. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Continual test-time adaptation (CTTA) adapts a source model to an unlabeled test stream whose distribution may change over time. Existing TTA methods often assess prediction reliability using confidence or entropy, which primarily reflect the model's self-certainty for the current sample. In CTTA, accumulated target observations can provide complementary evidence for correcting the source prediction, but this history may become misaligned as the target distribution changes. The key question is therefore not how much the correction differs from the source prediction, but whether and how strongly it should be applied. This paper proposes Gain-Aware INtervention (GAIN), a backpropagation-free CTTA framework guided by a simple principle: history proposes, gain decides. GAIN maintains compact target statistics to form a correction proposal and a posterior-predictive evaluator that accounts for estimation uncertainty. The resulting source-relative gain estimates the proposal's benefit and determines a sample-specific intervention strength along a continuous path through efficient one-dimensional optimization. Gain-controlled predictions then update the target statistics online, limiting the propagation of unreliable corrections, all without backpropagation, sample storage, or replay. Across five benchmarks, our method achieves strong predictive performance, with favorable accuracy--calibration--efficiency trade-offs in continual adaptation. On ImageNet-C, for example, GAIN achieves 61.9% accuracy with near-source calibration. It remains stable under diverse and challenging continual shifts while running 15.9x faster than a representative optimization-based CTTA baseline. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-30 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2609.36655 — 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 2026-10-07T18:04:24.726Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).