Deep learningDeep learning is related to a series of the state-of-the-art methods in contemporary artificial intelligence. In this chapter, our deep learning methods mainly include CNN and RNN models. In CNN models, YOLO models are especially emphasized, while in RNN models, we stress on transformer models for time series analysis along with LSTM. The transformer models are still large, active, and effective in our research projects, especially the diffusion transformer models for generative AI (GenAI). In this chapter, our focus is on vision transformer (ViTVision transformer (ViT)) for robotic scene understanding. The significance of this chapter is that the state-of-the-art knowledge in deep learning is mingled with the knowledge of robotic vision for developing autonomous systems.