Emotional regulation improves deep learning-based image classification

The study introduces Emotional Regulation, a novel framework that incorporates artificial subjective experience to enhance deep learning models by modeling emotion. It demonstrates improved performance in image classification tasks using ResNet and ViT architectures pre-trained on emotional datasets, achieving state-of-the-art results on CIFAR-10 and CIFAR-100 benchmarks. This approach highlights the significance of affective states in optimizing machine learning tasks, suggesting a new direction for emotion-augmented deep learning architectures.