This cluster of papers focuses on the development and improvement of visual question answering systems, image captioning techniques, and neural networks for understanding and generating descriptions of images and videos. The research involves semantic reasoning, multimodal fusion, scene graph generation, attention mechanisms, and deep learning approaches to bridge the gap between vision and language.