“Fed-DRL”: A Timeliness Optimization Method for Dynamic Data Acquisition System Based on Mobile Edge Computing

Figure 5

Deep reinforcement learning framework with an attention mechanism.