Metraipolis › archive › dc_9db428ec2874

CoRA-NAS: Coarse Ranking and Anchor-Residual Refinement for Neural Architecture Search — arXiv:2609.11884

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.11884
License (as archived)
CC0-1.0
Archived
2026-10-16T05:55:04.535Z
SHA-256 of the archived text
15f9425889673c93b4b70f2f3e7b01f1dcc103de982baac6658d9b44e4d4e488
Stable URL of this page
https://metraipolis.com/documento/dc_9db428ec2874

How to cite

CoRA-NAS: Coarse Ranking and Anchor-Residual Refinement for Neural Architecture Search — arXiv:2609.11884 — arXiv (en), https://arxiv.org/abs/2609.11884 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9db428ec2874 (Metraipolis archive dc_9db428ec2874, sha256 15f9425889673c93b4b70f2f3e7b01f1dcc103de982baac6658d9b44e4d4e488)
BibTeX
@misc{dc9db428ec2874,
  title = {CoRA-NAS: Coarse Ranking and Anchor-Residual Refinement for Neural Architecture Search — arXiv:2609.11884},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.11884},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9db428ec2874; sha256 15f9425889673c93b4b70f2f3e7b01f1dcc103de982baac6658d9b44e4d4e488},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.11884. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Zero-cost proxies rank architectures cheaply, but their reliability varies across search spaces. We introduce CoRA-NAS (COarse Ranking + Anchor-residual), a two-stage framework combining a static ranking prior with low-cost learning-curve refinement. CoRA-Rank aggregates capacity and structure-at-initialization proxies through an equal-weight log-rank consensus and a target-free consensus gate. CoRA-Refine samples anchors across this prior, extrapolates their early validation curves, and propagates a learned residual correction with an ExtraTrees model. The refinement uses approximately 1% of the cost of fully training the candidate set. Fully trained architecture-accuracy labels are not used to fit the ranker. One configuration is used across spaces, with space-specific architecture encodings. Across NAS-Bench-201, NAS-Bench-101, TransNAS-Bench-101, and NATS-SSS, CoRA-Refine achieves mean Spearman correlations of 0.946, 0.715, 0.786, and 0.894, respectively. Its worst-space correlation of 0.715 is the highest among the compared methods. On NAS-Bench-201/CIFAR-100, its selected architecture reaches 73.32% accuracy, near the reported ground-truth best of 73.37%. On the pure size space, refinement recovers the static prior's shortfall relative to parameter count, while remaining tied with the strongest capacity proxies within noise. The resulting framework combines cross-space ranking robustness with low-cost architecture selection.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-11 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.11884 — 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 2026-10-16T05:44:25.016Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).