Argument Mining Driven Analysis of Peer-Reviews Dataset

<p>Argument Mining in Scientific Reviews (AMSR)</p> <p>We release a new dataset of peer-reviews from different computer science conferences with annotated arguments, called AMSR (<strong>A</strong>rgument <strong>M</strong>ining in <strong>S</strong>cientific <strong>R</strong>eviews).<br> <br> The dataset has been crawled by the&nbsp;OpenReview&nbsp;platform (https://openreview.net/) and the OpenReviewCrawler (https://openreview-py.readthedocs.io/en/latest/getting data.html)<br> <br> From 12,135 collected papers and reviews, we sample 77 for the annotation.<br> We use a simple argumentation scheme,&nbsp;which distinguishes between non-arguments, supporting arguments, and attacking arguments, which we denote as NON/PRO/CON accordingly.</p>