The primary aim of this project is to obtain a representative subset of any given dataset, using which performance on any given task will be similar or approximately close to that of the entire dataset.

Please look into our Projects page to get more details.