Microsoft Document Aboutness Dataset
The Microsoft Document Aboutness Dataset consists of randomly sampled URLs (from a HEAD and TAIL distribution), all entities recognized in those documents, and a relevance assessment for each entity/URL pair as to whether or not…