SciDuet

SciDuet is the first publicaly available dataset for the challenging task of document2slides generation, The dataset integrated into GEM is the ACL portion of the whole dataset described in "aclanthology.org/2021.naacl-main.111.pdf". It contains the full Dev and Test sets, and a portion of the Train dataset. We additionally create a challenge dataset in which the slide titles do not match with the section headers of the corresponding paper.

Note that although we cannot release the whole training dataset due to copyright issues, researchers can still use our released data procurement code from github.com/IBM/document2slides to generate the training dataset from the online ICML/NeurIPS anthologies. In the released dataset, the original papers and slides (both are in PDF format) are carefully processed by a combination of PDF/Image processing tookits.