发表机构
Vilnius University; IRISA; Universite Bretagne Sud; European Commission Joint Research Center（维尔纽斯大学; 法国国家信息与自动化研究所雷恩分院（IRISA）; 南布列塔尼大学; 欧盟委员会联合研究中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究通过正则化进化搜索多层感知机架构，在保持任务、特征和划分不变的前提下，将湖泊叶绿素-a预测的留出AUC从0.790提升至0.820，准确率从0.733提升至0.748，参数减少26倍，模型仅1.6kB，适合机载筛选。
AI 中文摘要
在业务化地球观测中，带有专家设计的光谱特征的小型表格数据集是常态，而应用于这些数据集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藻华分类器——并询问在保持原始研究的任务、特征和湖泊级训练/测试划分不变的情况下，架构搜索能带来什么。使用正则化进化搜索扩展的多层感知机空间，并仅基于内部交叉验证AUC进行选择，我们发现网络将留出集AUC从0.790提高到0.820，准确率从0.733提高到0.748，同时仅使用409个可训练参数，比最强的手工设计参考模型少26倍。搜索收敛到一个一致的配方——一个狭窄的单层、RMS归一化、tanh激活、步长衰减的RMSprop和权重平均——这是实践者默认情况下不太可能达到的。所得模型大小为1.6kB，足够小以作为机载筛选触发器，这正是激励这项工作的场景。代码：此https URL。
英文摘要
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, $\tanh$ activation, step-decayed RMSprop and weight averaging -- that a practitioner would be unlikely to reach by default. At 1.6\,kB the resulting model is small enough to serve as an onboard screening trigger, which is the setting that motivates the work. Code: https://github.com/VU-AIML/automl4eo-bloom-nas.
CommentsAccepted at AutoML4EO 2026 (non-archival AutoML conference workshop). 4 pages + references. accepted-papers/" target="_blank" rel="noopener">https://automl4eo.org/accepted-papers/
