Automatic slide generation from scientific papers based on multimodal learning

Text Summarization for scientific papers is a very active topic of research with many state of the art solutions but its application in a Multimodal Setting, i.e., summarize the paper content in slides by considering also image content is a recent field of research with only few papers that tackle this task. The purpose of this work is to produce presentation slide for scientific papers by using a state of the art solution in the field of Extractive Text Summarization (only text) combined with a Multimodal model capable to learn visual concepts from images. Furthermore the aim of this report is to be an useful lecture for anyone that want to approach this task and help the production of future works.