[论文] Evolutionary Architecture Search for Chlorophyll-$a$ Prediction in Lak...

论文概要

研究领域: ML 作者: Kursat Komurcu, Linas Petkevicius 发布时间: 2026-10-07 arXiv: 2610.10496

中文摘要

在业务化地球观测中，带有专家设计光谱特征的小型表格数据集是常态，应用于其上的网络通常是手工设计的。我们重新审视了一个已发表的模型——一个Sentinel-2藻类水华分类器——在固定任务、特征和原始研究的湖泊级训练/测试划分的前提下，探究架构搜索能带来什么。用正则化进化搜索扩展的多层感知机空间，仅以内交叉验证AUC进行选择，我们找到的网络将留出AUC从0.790提升至0.820，准确率从0.733提升至0.748，同时仅使用409个可训练参数——比最强的手工设计参考少26倍。搜索收敛到一个一致的配方：单个窄层、RMS归一化、tanh激活、阶梯衰减RMSprop和权重平均——实践者不太可能默认想到这个方案。模型仅1.6kB，小到足以作为机载筛查触发器——这正是本工作的应用场景。代码：https://github.com/VU-AIML/automl4eo-bloom-nas

原文摘要

Small tabular datasets with expert-designed spectral features are the norm in operational Earth observation, and the networks applied to them are typically hand-designed. We revisit one such published model -- a Sentinel-2 algal bloom classifier -- and ask what architecture search adds, holding the task, the features and the lake-level train/test split of the original study fixed. Searching an extended multilayer-perceptron space with regularized evolution, and selecting on inner-cross-validation AUC only, we find networks that improve held-out AUC from 0.790 to 0.820 and accuracy from 0.733 to 0.748 while using 409 trainable parameters, 26 times fewer than the strongest hand-designed reference. The search converges on a consistent recipe -- a single narrow layer, RMS normalisation, $\tanh...

#论文 #arXiv #ML #小凯