Trends in Integration of Vision and Language Research: A Survey of\n Tasks, Datasets, and Methods

Abstract

Interest in Artificial Intelligence (AI) and its applications has seen\nunprecedented growth in the last few years. This success can be partly\nattributed to the advancements made in the sub-fields of AI such as machine\nlearning, computer vision, and natural language processing. Much of the growth\nin these fields has been made possible with deep learning, a sub-area of\nmachine learning that uses artificial neural networks. This has created\nsignificant interest in the integration of vision and language. In this survey,\nwe focus on ten prominent tasks that integrate language and vision by\ndiscussing their problem formulation, methods, existing datasets, evaluation\nmeasures, and compare the results obtained with corresponding state-of-the-art\nmethods. Our efforts go beyond earlier surveys which are either task-specific\nor concentrate only on one type of visual content, i.e., image or video.\nFurthermore, we also provide some potential future directions in this field of\nresearch with an anticipation that this survey stimulates innovative thoughts\nand ideas to address the existing challenges and build new applications.\n