Loss-Guided Model Sharing and Local Learning Correction in Decentralized Federated Learning for Crop Disease Classification

2025-05-29

AI Summary

To address data privacy and statistical heterogeneity in agricultural plant disease image classification, this paper proposes a fully decentralized federated learning framework that eliminates the central server and enables peer-to-peer collaborative training. Methodologically, it introduces two novel mechanisms grounded in validation loss: (1) an adaptive model-sharing strategy and (2) local loss-weighted correction, jointly optimizing inter-node knowledge exchange and local model updates. Furthermore, it integrates ResNet50, VGG16, and ViT-B16 into a unified architecture and designs a loss-aware adaptive weighting function to harmonize heterogeneous model contributions. Evaluated on the PlantVillage dataset, the approach achieves significantly higher accuracy and faster convergence than standard federated learning baselines, while demonstrating robustness under non-IID data distributions. Its serverless design and privacy-preserving nature make it particularly suitable for edge-deployed and privacy-sensitive agricultural applications.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Crop disease detection and classification is a critical challenge in agriculture, with major implications for productivity, food security, and environmental sustainability. While deep learning models such as CNN and ViT have shown excellent performance in classifying plant diseases from images, their large-scale deployment is often limited by data privacy concerns. Federated Learning (FL) addresses this issue, but centralized FL remains vulnerable to single-point failures and scalability limits. In this paper, we introduce a novel Decentralized Federated Learning (DFL) framework that uses validation loss (Loss_val) both to guide model sharing between peers and to correct local training via an adaptive loss function controlled by weighting parameter. We conduct extensive experiments using PlantVillage datasets with three deep learning architectures (ResNet50, VGG16, and ViT_B16), analyzing the impact of weighting parameter, the number of shared models, the number of clients, and the use of Loss_val versus Loss_train of other clients. Results demonstrate that our DFL approach not only improves accuracy and convergence speed, but also ensures better generalization and robustness across heterogeneous data environments making it particularly well-suited for privacy-preserving agricultural applications.

Problem

• Decentralized federated learning for crop disease classification
• Addressing data privacy in deep learning deployment
• Improving accuracy and robustness in heterogeneous data environments

Innovation

• Decentralized Federated Learning for privacy
• Loss_val guides model sharing adaptively
• Local learning correction via weighted loss