PaperRobot is a research system for automated scientific paper generation published at ACL 2019. It uses an incremental approach to produce paper sections—titles, abstracts, conclusions, and future work—from existing scientific literature. The model employs attention mechanisms and memory networks with PyTorch, trained on PubMed datasets containing over 875,000 training pairs for title-to-abstract and abstract-to-conclusion generation tasks.