Cover story: Zhang SX, Li H, Fan CZ et al. Adaptive decentralized AI scheme for signal recognition of distributed sensor systems. Opto-ElectronAdv 7, 240119 (2024).

Achieving large-scale security monitoring and precise threat identification is crucial for applications such as natural disaster early warning and infrastructure protection. Distributed sensing systems (DSS), by deploying large-scale sensor arrays in monitoring areas, enable comprehensive data acquisition about monitoring objectives. Traditionally, artificial intelligence (AI) model is trained using labeled data from one region (source region) of DSS, and then directly applied to other regions (target regions). However, this single "cross-region" model often struggles to handle significant differences between different regions, lacking the “brain-like” ability to adapt to unique regional characteristics. Recently, the research team of iF-Lab at Huazhong University of Science and Technology introduced an Adaptive Decentralized Artificial Intelligence (ADAI) scheme, designed to remodel the DSS "brain" in different regions. By leveraging unlabeled data from each target region, ADAI fine-tunes the model from source region, to generate adaptive cross-region model for each region. In the process of model finetuning, ADAI scheme establishes data relationships among different regions by measuring the feature distance between the source and target regions, and adapts the cross-region model to the target region by reducing the feature distance. In the application of intrusion signal recognition in distributed optical fiber sensing, the ADAI-enhanced cross-region model achieved average accuracy increases of 33.2% and 73% in two target regions, respectively, compared to baseline models. This method demonstrates a new paradigm of integrating AI with DSS, holding significant potential in threat identification of large-scale monitoring applications.