SSRCNN: A Semi-Supervised Learning Framework for Signal Recognition

Due to the emergence of deep learning, signal recognition has made great strides in performance improvement. The success of most deep learning methods relies on the accessibility of abundant labeled training data. However, the annotation of signals is quite expensive, making it challenging to train deep learning models substantially. This calls for the development of semi-supervised learning (SSL) method to fully utilize the unlabeled data to assist the training of deep learning models. To achieve this goal, three types of loss functions, tailored to the task of SLL-based signal recognition, are carefully designed in this paper. Together with a carefully selected neural network structure, the proposed SSL method can effectively extract the information from unlabeled training data and thus overcome the difficulty of insufficient training. Extensive numerical results using open source datasets are presented to show the superior performance of the proposed SSL method.