基于稀疏搜索历史与零成本代理的跨任务神经网络架构搜索方法
Searching with Sparse Histories: Cross-Task Architecture Search from Incomplete Search Logs and Zero-Cost Proxies
AI 导读
一项 arXiv 研究提出一种风险稳健的跨任务神经架构搜索方法，通过 champion-based 初始化与 joint pairwise-ranker 预训练，将稀疏源搜索日志与零成本代理结合。
正文 · 原文
Abstract:Neural architecture search (NAS) is expensive, but cross-task transfer can reduce costs by reusing knowledge from previous tasks. However, existing methods struggle with sparse search histories and task-specific transfer failures. We propose a method, a risk-robust strategy that combines sparse source logs with zero-cost proxies through champion-based initialization and joint pairwise-ranker pretraining. On TransNAS-Bench-101, the proposal mitigates severe worst-case failures, reducing worst-10% ranks from 290 to 43 for Micro autoencoding and from 465 to 45 for Macro room layout. Across 13 cell-space targets, it outperforms a proxy-only ranker on 10 targets in both mean and worst-case performance, while maintaining competitive average performance.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Neural and Evolutionary Computing (cs.NE) |
| Cite as: | arXiv:2610.32179 [cs.LG] |
| (or arXiv:2610.32179v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.32179 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jun Yu [view email]
[v1]
Fri, 23 Oct 2026 05:25:58 UTC (139 KB)
来源：arXiv cs.AI · arxiv.org