OpenD5 is a a meta-dataset which aggregates 675 open-ended problems ranging across business, social sciences, humanities, machine learning, and health, and uses a set of unified evaluation metrics: validity, relevance, novelty, and significance. It is designed for the new task, D5, that automatically discovers differences between two large corpora in a goal-driven way.
Source: Goal Driven Discovery of Distributional Differences via Language DescriptionsPaper | Code | Results | Date | Stars |
---|