Explainable Semantic Federated Learning Enabled Industrial Edge Network for Fire Surveillance

TL;DR

This paper tackles spectrum- and privacy-constrained fire surveillance in IIoT edge networks by introducing an Industrial Edge Semantic Network (IESN) that uses semantic communication to transmit compact semantic features. It develops eXplainable Semantic Federated Learning (XSFL), which combines privacy-preserving Federated Learning with an Adaptive Clients Training (ACT) strategy to tailor models to heterogeneous devices, and an Explainable SC (ESC) mechanism to render the learned semantics interpretable via gradient-based heatmaps. The proposed framework demonstrates that ACT improves training efficiency and model accuracy across devices, while ESC provides visual explainability that increases trust in the semantic predictions. Overall, IESN with XSFL reduces communication overhead, adapts to resource variance, and enhances explainability, offering a practical approach to robust, real-time fire surveillance at the industrial edge.

Abstract

In fire surveillance, Industrial Internet of Things (IIoT) devices require transmitting large monitoring data frequently, which leads to huge consumption of spectrum resources. Hence, we propose an Industrial Edge Semantic Network (IESN) to allow IIoT devices to send warnings through Semantic communication (SC). Thus, we should consider (1) Data privacy and security. (2) SC model adaptation for heterogeneous devices. (3) Explainability of semantics. Therefore, first, we present an eXplainable Semantic Federated Learning (XSFL) to train the SC model, thus ensuring data privacy and security. Then, we present an Adaptive Client Training (ACT) strategy to provide a specific SC model for each device according to its Fisher information matrix, thus overcoming the heterogeneity. Next, an Explainable SC (ESC) mechanism is designed, which introduces a leakyReLU-based activation mapping to explain the relationship between the extracted semantics and monitoring data. Finally, simulation results demonstrate the effectiveness of XSFL.

Figures (6)

• Figure 1: The illustration of IESN for fire surveillance. (a) Distributed SC system over the wireless network. (b) The architecture of the SC model.
• Figure 2: The illustration of the XSFL framework. (a) ACT strategy; (b) ESC mechanism.
• Figure 3: Model performance comparison among several FL schemes.
• Figure 4: The average training delay of each IIoT device using several FL schemes.
• Figure 5: The values of the optimization objective for different FL schemes.