发表机构
Chung Yuan Christian University（中原大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
AdapterMoE是一种两阶段硬路由MoE架构，用于多作物病害识别，通过RouterHead、双门限分布外模块及带温度缩放的作物专属Adapter，避免专家崩溃，降低训练成本，实现稳定弃权与增量学习，在PlantVillage数据集上表现优异。
AI 中文摘要
及时的作物病害识别对粮食安全至关重要。多作物识别适合采用混合专家（Mixture-of-Experts, MoE）架构，但传统的软路由MoE会在端到端训练中自由学习作物分配，导致少数专家主导（专家崩溃），且与作物无语义对应关系，还面临再训练成本高、非目标输入的弃权不稳定、准确率上限饱和等问题。我们将目标从准确率转向部署成本、扩展灵活性与弃权稳定性之间的权衡，采用确定性硬路由，提出AdapterMoE：RouterHead对作物进行分类，并通过最大Softmax概率阈值弃权非目标作物，采用双门限能量+KNN分布外模块捕捉分布偏移的输入；在冻结的EfficientNet-B0骨干之上，为每种作物配备5个Adapter以区分病害，每个Adapter通过温度缩放进行校准。由于专家在数据层面硬隔离，该设计避免了专家崩溃，并提供add_crop接口用于局部、逐作物更新，而非全量再训练。在PlantVillage数据集（5种作物、26个类别）上，经过公平的五系统对比，AdapterMoE的准确率与最优基线无统计学差异（Macro-F1在0.24个点的范围内），同时将训练成本降至全网络基线的约9%，可扩展至新作物。
英文摘要
Timely crop-disease identification is critical to food security. Multi-crop recognition suits Mixture-of-Experts (MoE), but conventional soft-routing MoE learns crop assignment freely end-to-end, letting a few experts dominate (expert collapse) with no semantic correspondence to crops, and facing high retraining costs, unstable rejection of non-target inputs, and a saturated accuracy ceiling. We shift the objective from accuracy toward a trade-off among deployment cost, scaling flexibility, and rejection stability, using deterministic hard routing. We propose AdapterMoE: a RouterHead classifies the crop and rejects non-target crops via a Maximum Softmax Probability threshold, with a dual-gate Energy+KNN out-of-distribution module catching distribution-shifted inputs; five per-crop Adapters atop a frozen EfficientNet-B0 backbone discriminate diseases, each calibrated via Temperature Scaling. Because experts are hard-isolated at the data level, the design avoids expert collapse and exposes an add_crop interface for local, per-crop updates instead of full retraining. On PlantVillage (5 crops, 26 classes), across a fair five-system comparison, AdapterMoE attains accuracy statistically indistinguishable from the best baselines (Macro-F1 within a 0.24-point band) while cutting training cost to about 9% of full-network baselines, expanding to a new crop in
Comments17 pages, 7 figures, 14 tables
