Meta Supervised Contrastive Learning for Few-Shot Open-Set Modulation Classification With Signal Constellation
Jikui Zhao, Huaxia Wang, Shengliang Peng, Yudong Yao
Abstract
In this work, we introduce a method for few-shot open-set modulation classification utilizing signal constellation diagrams, based on a Meta Supervised Contrastive Learning (MSCL) algorithm. MSCL combines the strengths of supervised contrastive learning and meta-learning to effectively amplify inter-class distinctions and reinforce intra-class compactness. The experimental results demonstrate that MSCL exhibits superior performance in both few-shot and open-set Automatic Modulation Classification (AMC) recognition. Code available at: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/jikuizhao/MSCL</uri>
Topics & Concepts
ConstellationComputer scienceModulation (music)SIGNAL (programming language)Set (abstract data type)Artificial intelligenceConstellation diagramPattern recognition (psychology)Speech recognitionTelecommunicationsDecoding methodsPhysicsBit error rateProgramming languageAstronomyAcousticsUltrasonics and Acoustic Wave PropagationWireless Signal Modulation ClassificationGeophysical Methods and Applications