Leveraging Contrastive Learning and Self-Training for Multimodal Emotion Recognition with Limited Labeled Samples | Read Paper on Bytez