Cite as and citations for datasets

[chiffa]

Hello there!

We would like refer to your repository in an academic article. Right now we are citing the repository by URL, but if you would like a more specific citation, a bibtex entry at the end of the readme would allow us to cite the the repository properly.

Similarly, to we are lacking as of now the citation for some of the datasets used. Based on the google drive files, the coco is a subset of MS COCO dataset, but there is no visible reference for the emnlp news dataset, which I assume is a news excerpt dataset recovered from an article published in emnlp? If this is correct, could you provide a reference to it?