arXiv：cs.LG（机器学习，全量分类）· Pascal Janetzky, Yuxin Wang, Michael Klar, Stefan Feuerriegel·· 15 小时前AI 评分32
PPNAS：面向神经架构搜索的预测驱动推理方法
Prediction-powered Neural Architecture Search
AI 导读
研究者提出 PPNAS，一种将预测驱动推理（PPI）用于神经架构搜索（NAS）的新方法，融合少量有性能标签的架构与大量含零成本代理（ZCP）信息的架构。PPNAS 利用 ZCP 提供的序数信息构建成对排序监督，并由 PPI 消除 ZCP 排序与真实性能排序之间的系统性偏差。在端到端基于预测器的 NAS 中，该方法在有限评估预算下达到 SOTA，作者称其为首个面向标签高效 NAS 的预测驱动方法。
正文
Abstract:Evaluating candidate architectures in neural architecture search (NAS) faces an inherent trade-off: on the one hand, reliable performance labels are limited because training and evaluating architectures is costly; on the other hand, zero-cost proxies (ZCPs) are cheap to compute at large scale but can be noisy. Yet, how to effectively combine these two sources of supervision remains unclear. In this paper, we propose PPNAS, a novel prediction-powered inference (PPI) approach for NAS. PPNAS fuses (1) a small set of architectures with observed performance labels and (2) a large set of architectures with ZCP information. To combine these two sources of supervision, PPNAS exploits the ordinal information provided by ZCPs to construct additional pairwise ranking supervision, while PPI debiases systematic discrepancies between ZCP-based and true performance rankings. We evaluate PPNAS in end-to-end predictor-based NAS, where it achieves state-of-the-art under limited evaluation budgets. To the best of our knowledge, PPNAS is the first prediction-powered approach for label-efficient NAS.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.01317 [cs.LG] |
| (or arXiv:2610.01317v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.01317 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Pascal Janetzky [view email]
[v1]
Thu, 1 Oct 2026 08:47:36 UTC (838 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org