SOTA?

Tracking the State-of-the-Art in Scholarly Publications

Dataset

The SOTA shared task comprises a dataset of Artificial Intelligence scholarly articles annotated with their (Task, Dataset, Metric, Score) tuples where applicable. The annotated training dataset is provided in the form of folders each consisting of two files: 1) scholarly articles in LaTeX format: and 2) an annotations file (Task, Dataset, Metric, Score) tuple annotations as a JSON dictionary for articles reporting leaderboards, otherwise a file with the string "unanswerable" for files not reporting leaderboards.

The SOTA? shared task Training data

Please visit https://github.com/jd-coderepos/sota/ to download the dataset.

Page updated

Google Sites

Report abuse