Research Article
Semi-supervised Learning for Automatic Modulation Recognition Using Haar Time–Frequency Mask and Positional–Spatial Attention
Algorithm 1
The Training Procedure of HTF-PSA-SSL.
Require:: student model with trainable parameters | Require:: teacher model with parameters equal to moving average of | Require:: labeled samples set | Require:: unlabeled samples set | Require:: learning rate of student model | Require:: rate of moving average | Require:: weight of unlabeled loss | Require:: Gaussian ramp–up curve function | Require:: batch size of labeled data | Require:: batch size of unlabeled data | Require:. | 1: for t = do | 2: Sample | 3: Calculate via Equation (8). | 4: Sample , | 5: Generate pseudo label , | 6: | 7: Calculate via Equation (10). | 8: | 9: Calculate via Equation (14). | 10: | 11: | 12: | 13: | 14: end for | 15: return, |
|